AGENT SKILL

fetch-tweets

Search and curate X/Twitter behind one selector - keyword, topic roundup, a single or tracked-account digest, an X list, or the AI-agent buzz preset - clustered into signal-scored sub-narratives.

License 许可
MIT
所属仓库
aeonfun/aeon
最近更新
2026-11-15
Stars
★ 793
Forks
308

01 前言 Front Matter

name
fetch-tweets
metadata
{ "var": "", "tags": [ "social" ], "title": "Fetch Tweets", "category": "basics", "requires": [ "TWITTER_API_KEY", "XAI_API_KEY?" ] }
description
Search and curate X/Twitter behind one selector - keyword, topic roundup, a single or tracked-account digest, an X list, or the AI-agent buzz preset - clustered into signal-scored sub-narratives.

03 SKILL.md 正文

<!-- autoresearch: variation B — sharper output via clustering + signal scoring + insight extraction. Merged HUB: absorbs tweet-digest, tweet-roundup, list-digest, refresh-x, agent-buzz behind a source: selector. -->

${var} — <source>:<arg> where <source> ∈ keyword | topic | account | list | agent-buzz. The <arg> is source-specific (a query, a topic, a handle, comma-separated list IDs, or an optional focus). If no source: prefix is given, the source is inferred from the shape of <arg> (see Source selector). Required for keyword and list; optional for topic, account, and agent-buzz.

Today is ${today}. This skill fetches X/Twitter content along one of five source axes and produces a curated digest — clustered by sub-narrative, ranked by signal, one insight per item — never a flat chronological dump.

Source selector

Parse ${var} into SOURCE and ARG before doing anything else.

Explicit form (recommended): <source>:<arg>

  • keyword:$SOL OR solana OR "solana network" - raw X search query, passed to the search verbatim (OR/AND honored; twitterapi advanced_search primary, Grok x_search fallback).
  • topic:brain-computer interfaces — a single topic roundup. topic: (empty arg) → resolve a topic list from MEMORY.md, then built-in defaults.
  • account:vitalikbuterin — one account's recent tweets. account: (empty arg) → digest every handle in memory/topics/tracked-accounts.yml.
  • list:1953536336675365173,1937207796270829766 — one or more numeric X list IDs. Append |<topic> for a topic booster: list:195...,193...|AI agents.
  • agent-buzz — the curated AI-agent-ecosystem preset. agent-buzz:MCP protocol prioritizes a project/topic within the preset.

Implicit form (back-compat with migrated bare-var configs): when ${var} has no recognized source: prefix, infer SOURCE in this order:

  1. ${var} is empty → topic (default multi-topic roundup).
  2. ${var} is all-digits, or comma-separated all-digits (optionally with a |<topic> suffix) → list.
  3. ${var} is @handle or matches ^[A-Za-z0-9_]{1,15}$ (a bare handle) → account.
  4. Anything else → keyword.

Note: agent-buzz has no distinct implicit shape (its arg looks like a keyword/topic), so it is only selectable via the explicit agent-buzz / agent-buzz:... prefix.

Once SOURCE and ARG are set, jump to the matching branch below. Only one branch runs per invocation.

Shared preamble (all branches)

  1. Read memory/MEMORY.md for context and the recent memory/logs/ (each branch specifies its lookback window — 2 or 3 days) to dedup already-reported tweets.

  2. Load the dedup set SEEN_TWEETS by unioning two sources:

    • The branch's persistent seen-file (per-mode path below), if it exists — read all URLs.
    • The branch's log lookback window — grep each memory/logs/*.md file in range for lines matching https://x.com/.

    Per-mode seen-files (kept at their legacy paths so dedup history survives the merge):

    mode seen-file log lookback
    keyword memory/fetch-tweets-seen.txt 3 days
    topic memory/tweet-roundup-seen.txt 3 days
    account (logs only — see branch) 2 days
    list memory/list-digest-seen.txt 2 days
    agent-buzz (logs only — 3-day status/<id> set) 3 days
  3. Formatting invariants shared by every branch's notification:

    • Use x.com/handle (never @handle) so Telegram doesn't ping/tag users. (Exception: the account-digest and agent-buzz formats below historically use @handle in-body; keep their documented format but prefer x.com/handle when practical.)
    • Every surviving tweet gets a tappable Markdown link — [View](url) / [View tweet](url). If a URL is unavailable, drop the link and say "(link unavailable)".
    • Never fabricate engagement counts. Missing → 0, not a guess.
    • Notify only on signal. A legitimately empty or all-duplicate run logs its status and sends nothing.

Voice

Used by the account and agent-buzz branches for one-line takes/insights. If soul/SOUL.md and soul/STYLE.md are populated, read both and match the operator's voice. If they are empty templates or absent, write in a clear, direct, neutral tone — state what the tweet says, no hedging or editorializing beyond the tweet itself.


Branch: keyword (source:keyword)

Search X for tweets matching ARG and produce a curated digest grouped by sub-narrative.

Seen set: memory/fetch-tweets-seen.txt + last 3 days of logs (loaded in preamble).

  1. Build the search prompt. Pass ARG to Grok verbatim as the query — do NOT narrow it to a single angle; broad coverage is the goal. Ask for at least 15–20 candidate tweets (you'll cull to ~7–10). Always require explicit engagement counts (likes, retweets, replies) so ranking is data-driven.

  2. Fetch tweets. Record SOURCE_PATH=twitterapi|api|websearch for the log.

    Get the window dates first, each date as its own Bash call (see Shell shape below):

    date -u -d yesterday +%Y-%m-%d
    
    date -u +%Y-%m-%d
    

    The first prints FROM_DATE, the second TO_DATE.

    Path A - twitterapi.io (primary; see the Fetching (all branches) contract): structured X search, no prompt. Put the date window inside the query with since:. Pass ARG as the query verbatim (OR/AND honored) plus since:<FROM_DATE>, both written in literally:

    ./secretcurl -s -o /tmp/tw-keyword.json -w 'twitterapi http=%{http_code} bytes=%{size_download}\n' --max-time 60 -G "https://api.twitterapi.io/twitter/tweet/advanced_search" \
      --data-urlencode "query=<ARG> since:<FROM_DATE>" --data-urlencode "queryType=Latest" \
      -H "X-API-Key: {TWITTER_API_KEY}"
    

    On HTTP=200 with a non-empty .tweets[], parse and mark SOURCE_PATH=twitterapi:

    jq -r '.tweets[] | [.id, .author.userName, .text, .url, .createdAt, .likeCount, .retweetCount, .replyCount] | @tsv' /tmp/tw-keyword.json
    

    Paginate with &cursor=<next_cursor> (from .next_cursor while .has_next_page) only if you need more than the first page to reach ~15-20 candidates. Engagement counts are exact ground-truth; permalinks are the real .url.

    Path B - xAI Grok x_search (fallback) (only if Path A returned non-2xx / empty / timeout, or TWITTER_API_KEY is unset; see the contract - set the Bash tool timeout >=180000, capture the HTTP status):

    jq -n '{model:"grok-4.7", input:[{role:"user",content:"Search X for tweets about: <ARG>. Date range: <FROM_DATE> to <TO_DATE>. Return at least 15-20 candidate tweets - mix of high-engagement posts and smaller accounts that add a distinct angle. For each tweet include: @handle, the full text, date posted, exact engagement counts (likes, retweets, replies - never N/A; if unknown, say 0), and the direct link (https://x.com/handle/status/ID). Return as a numbered list."}], tools:[{type:"x_search"}]}' \
      | ./secretcurl -s -o /tmp/xai.json -w '%{http_code}' --max-time 150 -X POST "https://api.x.ai/v1/responses" \
      -H "Content-Type: application/json" \
      -H "Authorization: Bearer {XAI_API_KEY}" \
      --data-binary @-
    

    On HTTP=200, parse /tmp/xai.json with: jq -r '.output[] | select(.type == "message") | .content[] | select(.type == "output_text") | .text' and mark SOURCE_PATH=api.

    Path C - WebSearch (last resort) (only if Paths A and B both gave a non-2xx / empty / timeout, or no key is set): use the built-in WebSearch tool with site:x.com "<query terms>" after:${FROM_DATE}. Note at the top of the log the true reason (http-<code> / timeout / empty, never "unavailable" when a key was set) and "results compiled via WebSearch - quality lower than usual". WebSearch favours high-engagement older tweets - prioritise results dated within the last 48 hours. Mark SOURCE_PATH=websearch.

  3. Empty vs. error handling (distinguish):

    • Legitimate empty (0 tweets): log FETCH_TWEETS_EMPTY (source=${SOURCE_PATH}) and stop — no notification.
    • API/cache error (HTTP error, malformed JSON, all paths failed): log FETCH_TWEETS_ERROR (last_path=${SOURCE_PATH}, reason=...) and stop — no notification.
  4. Deduplicate each candidate URL against SEEN_TWEETS. If ALL are dupes: log FETCH_TWEETS_NO_NEW: all results already reported and stop — no notification.

  5. Curate (the core step): a. Cluster survivors into 2–4 sub-narratives by what they're claiming/discussing (e.g. for a token: "price action", "team announcement", "criticism/FUD", "ecosystem integration"). Name the angle, not the topic. b. Rank within each cluster by signal (not raw engagement): signal = likes + 2×retweets + replies, but demote pure replies, generic shilling, and near-duplicate paraphrases. Drop tweets with <5 total engagement unless they add a unique angle. c. Cap each cluster at 2–3 tweets, total 7–10. Quality over quantity — if only 5 pass, send 5. Don't pad. d. Extract the claim/signal per tweet — what's new or interesting, not a literal paraphrase. Bad: "User says token is going up." Good: "Calls out the team's silence on the postponed unlock — first major holder to do so publicly." e. Compute a one-line signal for the top of the notification — one observation about the shape of the conversation (e.g. "Sentiment split — 4 bullish on the launch, 3 critical of the unlock terms.").

  6. Save + update seen-file (see Log). Append each kept tweet URL (one per line) to memory/fetch-tweets-seen.txt (create if missing).

  7. Notify via ./notify with the clustered output:

    *Top Tweets - ${ARG} (${today})*
    _${signal_one_liner}_
    
    *${cluster_1_name}*
    1. x.com/handle - [insight summary]
    Likes: X | RTs: Y | Replies: Z
    [View tweet](https://x.com/handle/status/ID)
    
    2. x.com/handle - [insight summary]
    Likes: X | RTs: Y | Replies: Z
    [View tweet](https://x.com/handle/status/ID)
    
    *${cluster_2_name}*
    3. x.com/handle - [insight summary]
    ...
    

    The signal one-liner is italic (_..._) directly under the title; cluster headers are *bold*.

Status codes: FETCH_TWEETS_OK (notified) | FETCH_TWEETS_EMPTY | FETCH_TWEETS_ERROR | FETCH_TWEETS_NO_NEW.


Branch: topic (source:topic)

Gist of the latest X chatter on one or more configurable topics.

Seen set: memory/tweet-roundup-seen.txt + last 3 days of logs.

  1. Resolve the topic list (priority order):

    1. ARG set → TOPICS=("$ARG") (single-topic mode).
    2. Else if MEMORY.md has a ## Tweet Roundup Topics section → use its bulleted lines, one query per line.
    3. Else built-in defaults:
      • artificial intelligence OR AI agents OR LLM
      • crypto OR bitcoin OR DeFi
      • technology OR startups OR open source
  2. Fetch per topic - track SOURCE ∈ {twitterapi, api, websearch, failed} per topic.

    Get the window dates first, each date as its own Bash call (see Shell shape below):

    date -u -d yesterday +%Y-%m-%d
    
    date -u +%Y-%m-%d
    

    The first prints FROM_DATE, the second TO_DATE.

    Path A - twitterapi.io (primary): for each topic, one call with the topic as the query and the window inside it via since:, both written in literally:

    ./secretcurl -s -o /tmp/tw-topic.json -w 'twitterapi http=%{http_code} bytes=%{size_download}\n' --max-time 60 -G "https://api.twitterapi.io/twitter/tweet/advanced_search" \
      --data-urlencode "query=<TOPIC> since:<FROM_DATE>" --data-urlencode "queryType=Latest" \
      -H "X-API-Key: {TWITTER_API_KEY}"
    

    On HTTP=200 with a non-empty .tweets[], parse and mark SOURCE=twitterapi. Extract each tweet's handle (.author.userName), text, engagement counts (.likeCount/.retweetCount/.replyCount), and permalink (.url); take up to the first 8 substantive tweets:

    jq -r '.tweets[] | [.id, .author.userName, .text, .url, .createdAt, .likeCount, .retweetCount, .replyCount] | @tsv' /tmp/tw-topic.json
    

    Path B - xAI Grok x_search (fallback) (only if Path A returned non-2xx / empty / timeout, or TWITTER_API_KEY unset): for each topic, call Grok's x_search.

    jq -n '{model:"grok-4.7", input:[{role:"user",content:"Search X for recent tweets about: <TOPIC>. Date range: <FROM_DATE> to <TO_DATE>. Return up to 8 substantive tweets. For each: @handle, full text, date, exact engagement counts (likes, retweets, replies; 0 if unknown), and the direct link https://x.com/handle/status/ID."}], tools:[{type:"x_search"}]}' \
      | ./secretcurl -s -o /tmp/xai-topic-out.json -X POST "https://api.x.ai/v1/responses" \
      -H "Content-Type: application/json" \
      -H "Authorization: Bearer {XAI_API_KEY}" \
      --data-binary @-
    

    Parse with the standard jq extractor. If it yields text, SOURCE=api. Extract each tweet's @handle, text, engagement counts, and permalink.

    Path C - WebSearch (last resort) (only if Paths A and B both error/empty, or no key set): site:x.com "<topic keywords>" after:<YESTERDAY>. Always include the word "today" and ${today} to force fresh results. Discard any result whose visible date is older than 48h. Collect up to 5 candidates per topic. Mark SOURCE=websearch. If all paths return nothing, mark SOURCE=failed.

  3. Score and filter. Require: a known @handle; a https://x.com/<handle>/status/<id> URL (if missing, keep but mark "link unavailable"); posted within 48h; URL not in SEEN_TWEETS. Compute signal_score = likes + 2×retweets + replies (on WebSearch path with no counts, use result rank as a weak proxy). Demote −50%: replies to a parent tweet; near-duplicates of a higher-scoring tweet (>70% text overlap or same linked URL).

  4. Curate per topic:

    • 0 survivors → drop the topic. Do NOT pad.
    • 1–3 survivors → list ranked by signal_score, highest first.
    • 4+ survivors → group into 2–3 sub-narratives (shared keywords/entity/claim); label each, surface the top-1 tweet per narrative as exemplar. Write an insight per reported tweet (what it asserts/reveals, not a headline paraphrase). Write a one-line conversation shape per topic ("bullish momentum, dissenters quiet", "split opinion on X's launch", "single story dominating — Y").
  5. Notify. If every topic dropped: log TWEET_ROUNDUP_EMPTY and stop — no notify. Otherwise send via ./notify (≤4000 chars):

    *Tweet Roundup - ${today}*
    _Source: api:X websearch:Y failed:Z_
    
    *[Topic 1]* - _conversation shape_
    - x.com/handle - insight (signal: 12.3k) [View](https://x.com/handle/status/ID)
    - x.com/handle - insight (signal: 4.1k) [View](https://x.com/handle/status/ID)
    
    *[Topic 2]* - _conversation shape_
    - x.com/handle - insight (signal: 8k) [View](https://x.com/handle/status/ID)
    

    Show signal: <score> only when engagement counts were available (api path); omit silently on WebSearch.

  6. Persist + log (see Log). Append each reported URL (one per line) to memory/tweet-roundup-seen.txt (create if missing).

Constraints: never notify an empty roundup (silence beats filler); never @handle anyone; never report a URL already in SEEN_TWEETS. Status codes: TWEET_ROUNDUP_OK | TWEET_ROUNDUP_EMPTY.


Branch: account (source:account)

Two sub-modes: single handle (decision-ready gist of one account) vs. all tracked accounts (theme-grouped digest of a watchlist). Choose by ARG.

Seen set: last 2 days of logs — extract every https://x.com/ URL under a prior ### fetch-tweets account entry into SEEN_URLS.

account — single handle (ARG is one @handle)

  1. Normalize ARG. Strip leading @, https://x.com/, https://twitter.com/, https://nitter.net/, trailing slash / /status/.... Lowercase. Reject if empty, contains whitespace, or >15 chars. On reject → REFRESH_X_NO_VAR: send ./notify "fetch-tweets: REFRESH_X_NO_VAR — set an X handle" and exit 0. Store the cleaned handle as ACCOUNT.

  2. Load tweets:

    • Path A - twitterapi.io (primary): pull this account's recent timeline directly, then filter to the last 2 days client-side (no server-side date filter on this endpoint): First date -u -d '2 days ago' +%Y-%m-%d as its own call (prints FROM_2D), then one call with the handle written in literally:
      ./secretcurl -s -o /tmp/tw-account.json -w 'twitterapi http=%{http_code} bytes=%{size_download}\n' --max-time 60 -G "https://api.twitterapi.io/twitter/user/last_tweets" \
        --data-urlencode "userName=<ACCOUNT>" -H "X-API-Key: {TWITTER_API_KEY}"
      
      On HTTP=200 with a non-empty .data.tweets[], parse and mark source=twitterapi. Filter on .createdAt (drop tweets older than the window) and use .isReply to tag reply vs original (this endpoint has no quote flag - treat non-reply as original; if you need reliable quote detection fall back to Path B). One jq per call, FROM_2D written in literally:
      jq -r '.data.tweets[] | select((try (.createdAt | strptime("%a %b %d %H:%M:%S %z %Y") | strftime("%Y-%m-%d")) catch "0000-00-00") >= "<FROM_2D>") | [.id, .author.userName, .text, .url, .createdAt, .likeCount, .retweetCount, .replyCount, .isReply] | @tsv' /tmp/tw-account.json
      
    • Path B - xAI Grok x_search (fallback) (only if Path A returned non-2xx / empty / timeout, or TWITTER_API_KEY unset): search this account's recent tweets via Grok's x_search.
      jq -n '{model:"grok-4.7", input:[{role:"user",content:"Search X for the latest tweets, replies, and quote tweets from @<ACCOUNT> in the last 2 days. Return each with full text, timestamp, type (original|reply|quote), what it replies to/quotes if any, exact engagement counts (likes, retweets, replies; 0 if unknown), and the permalink https://x.com/<ACCOUNT>/status/ID. Skip retweets of others. Return chronological."}], tools:[{type:"x_search"}]}' \
        | ./secretcurl -m 30 -s -o /tmp/xai-account-out.json -X POST "https://api.x.ai/v1/responses" \
        -H "Content-Type: application/json" \
        -H "Authorization: Bearer {XAI_API_KEY}" \
        --data-binary @-
      
      Parse with the standard jq extractor. Record source=api.
    • Path C - WebFetch (last resort) (only if Paths A and B both failed / returned nothing): WebFetch https://x.com/${ACCOUNT} with prompt: "List every tweet, reply, and quote tweet visible on this profile with its full text, timestamp, engagement counts (likes/retweets/replies) if shown, and the permalink https://x.com/handle/status/ID. Return a chronological list." Record source=webfetch.
    • Degraded: if no key is set and WebFetch returns nothing → skip to step 8 with status REFRESH_X_NO_API_KEY (no key present) or REFRESH_X_ERROR (a key was set but all paths failed).
  3. Parse into structured tweets: url, text, timestamp, type (original/reply/quote), reply_to, quoted_text, likes, retweets, replies. Drop retweets of others. Missing counts → 0. Compute signal_score = likes + 2*retweets + replies − (3 if type=reply else 0).

  4. Dedup and gate: drop any tweet whose url is in SEEN_URLS (deduped_count). If fewer than 3 tweets survive AND no thread is detectable (step 5) → skip to step 8 with REFRESH_X_NO_NEW (everything deduped) or REFRESH_X_EMPTY (account posted nothing).

  5. Detect threads: a thread = 2+ tweets by ACCOUNT within 30 minutes where later tweets reply to earlier ones OR share ≥2 meaningful keywords with the opener. Thread tweets are atomic units regardless of individual score. Record {opener_url, tweet_count, combined_signal}.

  6. Cluster and extract insights: group survivors (threads = one unit) into 2–4 sub-narratives by topic overlap; if <2 emerge, use one cluster. Per cluster: Title (3–8 words), Top tweet(s) (1–3 excerpts ≤200 chars each, with permalink + engagement), Insight (one sentence — what the cluster reveals about the author's stance/claim/shift; not a paraphrase — if you can't beat paraphrase, drop the cluster). Per thread: a 1–2 sentence landing summary + opener URL.

  7. Write the verdict (pick exactly one) + a ≤20-word lede:

    Verdict When
    ANNOUNCEMENT launch, hire, policy, or product drop
    ARGUMENT majority signal from contrarian takes or fights
    BUILDING ships/code/tech-progress clusters dominate
    SHITPOST jokes, memes, low-stakes banter dominate
    CONTEXT mostly reacting to a news cycle, not driving one
    QUIET <3 originals and no thread
  8. Save gist (see Log). On empty/no-new/error/no-var statuses, write only the account header + status footer, skip cluster sections.

  9. Update MEMORY.md (conditional): only if a cluster carries an announcement, specific claim, named project, or stance shift — add one bullet under a ## Tracked X Accounts section (create if missing): - @ACCOUNT YYYY-MM-DD: [one-sentence claim] — [permalink]. No paraphrases/memes/generic opinions.

  10. Notify via ./notify. On REFRESH_X_OK:

    x refresh - @ACCOUNT ([VERDICT])
    [lede]
    top cluster: [title] - "[≤80 char excerpt]" ([likes]❤)
    [N tweets, T threads, K deduped]
    

    On REFRESH_X_EMPTY / REFRESH_X_NO_NEW: skip notify (write the log entry only). On REFRESH_X_NO_API_KEY / REFRESH_X_ERROR / REFRESH_X_NO_VAR: notify with the status code + a one-line hint (e.g. "fetch-tweets: REFRESH_X_NO_API_KEY — set XAI_API_KEY in workflow secrets").

Constraints: never fabricate engagement; never include a SEEN_URLS URL; an insight that only paraphrases is not an insight (drop the cluster); MEMORY.md updates are one line each. Status codes: REFRESH_X_OK | REFRESH_X_EMPTY | REFRESH_X_NO_NEW | REFRESH_X_NO_API_KEY | REFRESH_X_ERROR | REFRESH_X_NO_VAR.

account — all tracked accounts (ARG empty)

Use this to answer "what did these specific people post" across a watchlist.

  1. Read config memory/topics/tracked-accounts.yml. If missing or accounts: [] → log TWEET_DIGEST_NO_CONFIG and exit (no notification). Schema:

    accounts:
      - handle: vitalikbuterin
        why: ethereum core thinking      # optional — grouping/context label
      - handle: balajis
        why: macro + tech narratives
    
  2. Fetch recent tweets per account. For each handle:

    • Path A - twitterapi.io (primary): pull the handle's recent timeline, filter to the last 3 days client-side, take the 5 most substantive; drop retweets of others: Run date -u -d '3 days ago' +%Y-%m-%d once, as its own call (prints FROM_3D). Then per handle, one call with the handle written in literally:
      ./secretcurl -s -o /tmp/tw-acct1.json -w 'twitterapi http=%{http_code} bytes=%{size_download}\n' --max-time 60 -G "https://api.twitterapi.io/twitter/user/last_tweets" \
        --data-urlencode "userName=<HANDLE>" -H "X-API-Key: {TWITTER_API_KEY}"
      
      On HTTP=200 with a non-empty .data.tweets[], parse and mark source=twitterapi. One jq per call, FROM_3D written in literally; if you also want a count, jq '.data.tweets | length' /tmp/tw-acct1.json is its own call, never joined with ;:
      jq -r '.data.tweets[] | select((try (.createdAt | strptime("%a %b %d %H:%M:%S %z %Y") | strftime("%Y-%m-%d")) catch "0000-00-00") >= "<FROM_3D>") | [.id, .author.userName, .text, .url, .createdAt, .likeCount, .retweetCount, .replyCount] | @tsv' /tmp/tw-acct1.json
      
    • Path B - xAI Grok x_search (fallback) (only if Path A returned non-2xx / empty / timeout, or TWITTER_API_KEY unset):
      jq -n '{model:"grok-4.7", input:[{role:"user",content:"Search X for the latest tweets from:<HANDLE> in the last 3 days. Return the 5 most interesting or substantive tweets. For each: full text, date, direct link (https://x.com/<HANDLE>/status/ID). Skip retweets of others."}], tools:[{type:"x_search"}]}' \
        | ./secretcurl -m 30 -s -o /tmp/xai-acct1-out.json -X POST "https://api.x.ai/v1/responses" \
        -H "Content-Type: application/json" \
        -H "Authorization: Bearer {XAI_API_KEY}" \
        --data-binary @-
      
      Parse with the standard jq extractor. If neither TWITTER_API_KEY nor XAI_API_KEY is set, log TWEET_DIGEST_NO_KEY: skill requires TWITTER_API_KEY (or XAI_API_KEY) and exit (no notification). Dedup: drop any candidate URL already in SEEN_URLS (last 2 days of logs).
  3. Group by theme, not by account. Walk the full candidate set; identify 2–4 themes (e.g. "L2 design decisions", "macro / rates", "AI model releases", "regulation"). Each tweet maps to one theme; a why: label can seed theme naming for single-topic feeds.

  4. Write a one-sentence take per notable tweet — what the tweet says, not your opinion of it. Voice per the Voice section.

  5. Notify via ./notify:

    *Tweet Digest - ${today}*
    
    *Theme: <theme>*
    @handle: <one-sentence summary> - [link](url)
    @handle: <one-sentence summary> - [link](url)
    
    *Theme: <theme>*
    ...
    

    If no notable tweets across all accounts: log TWEET_DIGEST_OK and end (no notification).

Status codes: TWEET_DIGEST_OK (notified or clean) | TWEET_DIGEST_NO_CONFIG | TWEET_DIGEST_NO_KEY.


Branch: list (source:list)

Cross-list narrative resonance + signal-scored top tweets from tracked X lists in the past 24h. Lists are curator signal — the value is cross-list resonance + insight + a verdict, not a flat top-N-per-list dump.

Seen set: memory/list-digest-seen.txt + last 2 days of logs.

  1. Parse and validate ARG in your head, not in a shell script (an if/for block is refused):

    • ARG empty: log LIST_DIGEST_NO_CONFIG: var must contain at least one X list ID to today's memory/logs/<date>.md and stop.
    • IDS_PART = the text before the first |; TOPIC_FILTER = the text after it (empty when there is no |).
    • Split IDS_PART on commas into LIST_IDs. Any ID that is not all digits: log LIST_DIGEST_NO_CONFIG: invalid list ID '<id>' (must be numeric) and stop.

    If XAI_API_KEY is unset, fall back to Path B. If no path returns data, log LIST_DIGEST_NO_CONFIG: XAI_API_KEY required and stop without notifying.

  2. Fetch each list's top tweets (past 24h) - xAI primary, WebSearch fallback.

    Note: twitterapi.io has no list: operator (its advanced_search supports from: to: url: lang: #tag @mention since: until: only), so this branch keeps the xAI Grok x_search path as primary - it is the only path that can read an X list by ID. TWITTER_API_KEY does not apply to the list source. Path A - X.AI Responses API (primary). Get the window dates first, each date as its own Bash call: date -u -d yesterday +%Y-%m-%d (FROM_DATE) and date -u +%Y-%m-%d (TO_DATE). Then build the body and send it in one call (jq piped straight into ./secretcurl), with the list ID and both dates written in literally:

    jq -n '{model:"grok-4.7", input:[{role:"user",content:"Look at X list https://x.com/i/lists/<LIST_ID>. Step 1: report the list name and a one-line description. Step 2: identify the most engaging tweets posted by members of this list between <FROM_DATE> and <TO_DATE> UTC. Return the top 12 tweets ranked by engagement (likes, retweets, replies). For EACH tweet you MUST return: (a) @handle, (b) the full tweet text (not a paraphrase), (c) explicit engagement counts as separate fields - likes:N, retweets:N, replies:N, views:N if available, (d) the direct permalink in the form https://x.com/<handle>/status/<id>, (e) media type (image|video|none), (f) one-line context if it is a reply or quote tweet (who/what). Skip retweets of accounts NOT on this list. If a tweet has an image and you can analyze it, include a one-line image description."}], tools:[{type:"x_search", from_date:"<FROM_DATE>", to_date:"<TO_DATE>", enable_image_understanding:true}]}' \
      | ./secretcurl -s -o /tmp/xai-list-out.json --max-time 180 -X POST "https://api.x.ai/v1/responses" \
      -H "Content-Type: application/json" \
      -H "Authorization: Bearer {XAI_API_KEY}" \
      --data-binary @-
    

    Parse with the standard jq extractor. Path B — WebSearch fallback (only if XAI_API_KEY unset, OR Path A errors / returns nothing): site:x.com "i/lists/${LIST_ID}" OR list:${LIST_ID} after:${FROM_DATE}. Lower quality; mark this list's source as websearch. Per-list outcome: ok (≥3 tweets) | quiet (1–2) | empty (0, list found but no posts) | error (API/access failure — note reason).

  3. Build the candidate pool. Record per tweet {handle, text, likes, retweets, replies, views, url, list_ids_seen_on:[], list_names_seen_on:[], media, is_reply, is_quote}. Dedup by URL across lists — same tweet on multiple lists → merge records, keep both list_ids_seen_on and list_names_seen_on (cross-list appearance is a signal). Dedup against history — drop URLs in memory/list-digest-seen.txt or the last 2 days of logs.

  4. Score every candidate (natural-log engagement to stop one viral tweet dominating):

    base = ln(1+likes) + 2.0*ln(1+retweets) + 1.5*ln(1+replies)
    bonuses:
      +2.0  appeared on ≥2 distinct lists (cross-list resonance)
      +1.5  appeared on ≥3 distinct lists
      +1.0  topic_filter set AND tweet text/context matches (case-insensitive substring or obvious semantic match)
      +0.5  small-account-signal (≤25k followers per Grok's note OR no follower data + technical/insider content)
      +0.3  media is image OR video
    penalties:
      -1.0  is_reply AND replied-to NOT on any tracked list
      -0.5  pure link share with <10 words of original commentary
    score = base + sum(bonuses) - sum(penalties)
    
  5. Cluster into cross-list narratives when ALL hold: ≥2 tweets from ≥2 distinct lists; shared ≥2 substantive keywords/entities (proper nouns, project names, tickers, technical terms — ignore stop words); posted within the same 24h window. narrative score = sum of constituent tweet scores; narrative title ≤80 chars capturing what the cluster collectively says. Pick an anchor tweet (highest individual score) + up to 2 supporting. Cluster-count cap: if clustering yields <2 or >4 clusters, fall back to a flat ranked list with inline [cluster-name] labels (no "? Cross-list narratives" section).

  6. Compose the digest (cap 4000 chars): up to 3 narratives at top (by narrative score); then up to 5 standalone tweets per list (highest individual score, not already in a narrative); hard total cap 12 items — cut from the bottom of standalones. Insight discipline: every item needs a one-line so-what (implication, contrarian angle, missing number, deal-flow signal); a paraphrase must be rewritten. Quiet-list rule: if a list's top surviving tweet scores <2.0 (≈<8 likes raw), write a one-line "quiet day" for that list. Topic filter is a scoring booster (step 4), NOT a hard filter. Verdict line: one line at the very top capturing what today's lists collectively say.

  7. Send the notification via ./notify, verbatim format (x.com/handle, [label](url)):

    *List Digest - ${today}*
    
    [VERDICT LINE - one line, ≤140 chars, plain text]
    
    ? *Cross-list narratives*
    1. *[narrative title]* - appeared on [List A] + [List B]
       x.com/handle: [insight, not paraphrase] (♥ likes, ↻ rt) - [View](url)
       x.com/handle2: [insight] (♥ likes, ↻ rt) - [View](url)
    
    2. *[narrative title]* - appeared on [List A] + [List C]
       ...
    
    *[List Name 1]*
    - x.com/handle - [insight] (♥ likes, ↻ rt) - [View](url)
    - x.com/handle - [insight] (♥ likes, ↻ rt) - [View](url)
    
    *[List Name 2]*
    - quiet day
    
    ---
    sources: list1=ok | list2=quiet | list3=error(no-access)
    status: LIST_DIGEST_OK
    

    If cross-list narratives is empty, drop that whole section. If every list is quiet/empty, send a single-line "List Digest — ${today} — quiet across all tracked lists" instead of padding.

  8. Log and persist (see Log). Append every reported URL (one per line) to memory/list-digest-seen.txt (create if missing).

Exit taxonomy: LIST_DIGEST_NO_CONFIG (var empty/invalid OR no fetch path — log only) | LIST_DIGEST_EMPTY (every list 0 tweets OR all candidates already seen — log only) | LIST_DIGEST_PARTIAL (some lists succeeded/some failed — notify survivors, surface failures) | LIST_DIGEST_OK (≥1 fresh tweet — notify).


Branch: agent-buzz (source:agent-buzz)

A topic-filtered preset: a curated, narrative-aware read on what the AI-agent scene on X talked about in the last 24h. Curation, not aggregation — 6 high-signal tweets in 2 clusters beats 10 of mixed noise. ARG (optional) is a project/topic to prioritize.

Seen set: last 3 days of logs — extract every https://x.com/.../status/<id> already posted by this skill; treat those IDs as the dedup set.

  1. Fetch candidates: Get the window dates first, each date as its own Bash call (see Shell shape below):

    date -u -d '1 day ago' +%Y-%m-%d
    
    date -u +%Y-%m-%d
    

    The first prints FROM_DATE, the second TO_DATE.

    Path A - twitterapi.io (primary): structured search over the AI-agents conversation, window inside the query via since:, FROM_DATE written in literally:

    ./secretcurl -s -o /tmp/tw-buzz.json -w 'twitterapi http=%{http_code} bytes=%{size_download}\n' --max-time 60 -G "https://api.twitterapi.io/twitter/tweet/advanced_search" \
      --data-urlencode 'query=("AI agents" OR "autonomous agents" OR "agent framework" OR "agent frameworks" OR MCP OR "agent protocol" OR "agent benchmark") since:<FROM_DATE>' \
      --data-urlencode "queryType=Latest" -H "X-API-Key: {TWITTER_API_KEY}"
    

    On HTTP=200 with a non-empty .tweets[], parse and mark source=twitterapi. Collect up to ~40 candidates (paginate with &cursor=<next_cursor> while .has_next_page if the first page is thin):

    jq -r '.tweets[] | [.id, .author.userName, .author.isBlueVerified, .text, .url, .createdAt, .likeCount, .retweetCount, .replyCount] | @tsv' /tmp/tw-buzz.json
    

    Note: advanced_search does not return follower_count or role_guess. On this path treat both as null (the step-3 role multiplier and the follower-ratio skip-gates simply don't fire); use .author.isBlueVerified as a weak credibility proxy and lean on raw engagement. If ARG is set, append its terms to the query (AND) or issue a second constrained search and merge.

    Path B - xAI Grok x_search (fallback) (only if Path A returned non-2xx / empty / timeout, or TWITTER_API_KEY unset; the response for each tweet must include explicit engagement counts + follower count, or step 3 scoring can't run):

    jq -n '{model:"grok-4.7", input:[{role:"user",content:"Search X from <FROM_DATE> to <TO_DATE> for tweets in the AI-agents conversation: autonomous agents, agent frameworks, MCP / agent protocols, agent products, agent benchmarks, agent research papers. Return up to 40 candidates. For EACH candidate you MUST return: @handle, follower_count (integer or null), role_guess (builder|founder|researcher|investor|commentator|anon), one-line claim (what they actually said - not a paraphrase, the thesis), likes (int), retweets (int), replies (int), posted_at (ISO), direct_link (https://x.com/username/status/ID). Prefer builders/founders/researchers. Skip obvious engagement-farming threads (\"RT if you agree\", reply-guy pileons, giveaways)."}], tools:[{type:"x_search", from_date:"<FROM_DATE>", to_date:"<TO_DATE>"}]}' \
      | ./secretcurl -s -o /tmp/xai-buzz-out.json -X POST "https://api.x.ai/v1/responses" \
      -H "Content-Type: application/json" \
      -H "Authorization: Bearer {XAI_API_KEY}" \
      --data-binary @-
    

    Parse with the standard jq extractor. Record source=xai. Path C - WebSearch (last resort) (only if Paths A and B both error/empty, or no key set): forced-fresh query "AI agents twitter today ${today}" - discard anything >48h old, expect degraded metadata. Record source=websearch.

    If ARG is set, also issue a second call constrained to that topic with the same schema; merge results.

  2. Skip-gates (before clustering) — drop any candidate matching ANY:

    • Dup: status/<id> already in the 3-day dedup set.
    • Engagement-farming: poll threads, "bookmark this", "drop a ?", reply-guy pileons with <follower_count/10 likes.
    • Self-promo only: pure product shill with no claim/benchmark/datapoint. Launch tweets OK IF they include a concrete capability claim or number.
    • Staleness: posted_at older than 30h.
    • Anon + low engagement: role_guess=anon AND (likes+retweets) < 200.
  3. Signal scoring: signal = likes + 2*retweets + replies, then × 1.3 if role_guess ∈ {builder, founder, researcher}; × 0.7 if a pure hot-take with no concrete referent (no named project, number, paper, or bench); × 0.5 if near-duplicate of another survivor (keep the higher-scored one only).

  4. Narrative clustering: group survivors into 2–4 narrative clusters — a cluster is a shared thesis, not a keyword ("MCP vendor lock-in debate", not "MCP"). Name each ≤5 words. If one cluster holds >60% of tweets, split it. A tweet fitting no cluster is dropped unless its signal is top-3 overall. Target: 2–4 clusters, 2–3 tweets each, 6–9 total (strictly ≤10).

  5. Insight extraction — per tweet, a one-line insight (≤20 words): the actual claim/datapoint, not a paraphrase; if opinion, state what they're arguing against; if an announcement, state what's new vs. prior art (not "X launched"). Anti-hype lint — rewrite any insight containing: game-changing, revolutionary, mind-blowing, wild, huge, massive, unreal, insane, vague "AI agents are evolving", "the future of X".

  6. Conversation-shape lead — one opening sentence (≤25 words) naming what the conversation was actually about ("Mostly protocol debate — MCP vs. A2A — with two concrete launches on the side."). If you can't characterize it honestly in one sentence, the clustering is wrong — redo step 4.

  7. Notify via ./notify:

    *Agent Buzz - ${today}*
    _<conversation-shape one-liner>_
    
    **<Cluster 1 name>**
    • @handle - <insight>
      <link>
    • @handle - <insight>
      <link>
    
    **<Cluster 2 name>**
    • @handle - <insight>
      <link>
    
    <!-- _src: xai|websearch · candidates: N → kept: M_ -->
    

    Keep the footer — it's how future self-audits debug empty days. Never pad to hit 10. 6 good > 10 mid.

Status codes: AGENT_BUZZ_OK (≥1 cluster notified) | AGENT_BUZZ_EMPTY (fetch succeeded, nothing survived — send Agent Buzz — ${today}: quiet day, no survivors.) | AGENT_BUZZ_ERROR (all sources failed — notify Agent Buzz — ${today}: all sources failed (${error summary}). and log the per-source failure).


Log (all branches)

Append ONE entry per run to memory/logs/${today}.md under a single ### fetch-tweets heading (the health loop parses this shape). The first bullet is the discriminator naming the branch/mode that ran; the rest are branch-specific bullets. Always include the reported tweet URLs as bullets (for next-run dedup).

### fetch-tweets
- mode: <keyword|topic|account|list|agent-buzz>
- status: <STATUS_CODE for the branch that ran>
- source: <SOURCE_PATH / per-source counts / per-list outcome, as applicable>
- <branch-specific bullets - carry over each branch's fields:>
    - keyword:     signal one-liner; per-cluster URLs with `likes:N rts:N replies:N` + insight
    - topic:       `topics: [t1: N tweets, t2: 0 (dropped)]`; `source: cache:X websearch:Y failed:Z`
    - account(1):  Verdict + lede; Counts (N tweets, X orig/Y reply/Z quote, T threads, deduped K); Clusters; Threads; Vibe
    - account(all):themes covered; per-account tweet counts
    - list:        Lists tracked; Per-list `list1=ok(N) | list2=quiet(N) | list3=error`; Verdict; Narratives count
    - agent-buzz:  source used; candidates N → kept M; cluster names
- urls:
    - https://x.com/handle1/status/...
    - https://x.com/handle2/status/...

On empty/no-new/error/no-config statuses, write the ### fetch-tweets heading + mode: + status: bullets only (skip the detail sections) so skill-health still observes the run. After logging, update the branch's persistent seen-file where one exists (keyword / topic / list — see the seen-file table).

Output shape note

No chain consumes this skill's output as of this commit (no consume: [fetch-tweets] references). If a downstream chain step starts consuming it, emit a flat list of URLs before the clustered/branch output so consumers aren't broken by cluster or narrative headers.

Fetching (all branches)

Shell shape. Each Bash call is one allowed command (pipes are fine). The run refuses a call that sets a variable (FROM_DATE=$(date ...), HTTP=$(./secretcurl ...)), uses if / for / while, or joins commands with ;, && or ||, and each refusal costs a model turn. So get a date with a bare date call and write the printed value into later commands; write ARG, handles, topics and dates into each command literally (the <FROM_DATE>, <HANDLE> style placeholders in the snippets); run one jq per call; loop over handles or topics by making one call per item. The run also refuses any redirect into a file (> /tmp/x.json, cat > f) and any $name in a jq filter (--arg FROM ... '... $FROM'), in both tiers: write each value into the jq filter as a JSON string (>= "<FROM_2D>"; a " inside a value becomes \", and reword any '), and pipe a request body straight into ./secretcurl ... --data-binary @-.

Every branch fetches X in a three-tier cascade. Attempt each tier in order; only drop to the next on a real failure (non-2xx, empty body, or timeout), and always record the true reason.

Path A - twitterapi.io (primary). TWITTER_API_KEY is injected into your environment for this skill (declared in requires:). The primary fetch path in every branch (except list, which has no twitterapi list operator) is a direct curl to https://api.twitterapi.io/twitter/... with the X-API-Key: {TWITTER_API_KEY} header. It is fast (~700ms), returns structured ground-truth JSON with exact engagement counts and real permalinks (.url) - so there is no fabrication risk and no prompt to coax. Two endpoints:

  • Search: GET /tweet/advanced_search?query=<q>&queryType=Latest -> {tweets:[...], has_next_page, next_cursor}. Put date windows inside the query with X operators since:YYYY-MM-DD until:YYYY-MM-DD; other operators: from: to: url: lang: #tag @mention, plus OR/AND and quotes. Tweets are under .tweets[]. (There is no list: operator - see the list branch.)
  • Timeline: GET /user/last_tweets?userName=<handle> -> {status:"success", data:{tweets:[...]}, has_next_page, next_cursor}. No server-side date filter - filter client-side on .createdAt or paginate with &cursor=<next_cursor>. Tweets are under .data.tweets[].

Fields: id, text, url, createdAt, likeCount, retweetCount, replyCount, quoteCount, viewCount, author.userName, author.name, author.isBlueVerified (last_tweets also isReply, author.followers). Capture the status with -w 'twitterapi http=%{http_code} bytes=%{size_download}\n' (curl prints it, so no HTTP=$(...) wrapper) and read the http=<code> line before deciding anything. HTTP=200 with a non-empty tweet array -> use it (SOURCE_PATH=twitterapi). No prefetch, no cache - just make the call.

Path B - xAI Grok x_search (fallback). Only when Path A fails (non-2xx / empty / timeout) or TWITTER_API_KEY is unset. Direct curl to https://api.x.ai/v1/responses with Authorization: Bearer {XAI_API_KEY} using Grok's x_search tool. The x_search call typically takes 30-120s (it searches X live), so set the Bash tool timeout to at least 180000 (180s) and add --max-time 150 to the curl so it fails cleanly rather than hanging. A slow curl is not a missing key. Build the JSON body with jq -n (prompt and dates written into the filter literally) and pipe it straight into ./secretcurl in the same call (its body flag reads stdin) - every ./secretcurl command must be 100% literal (no shell variables, or the permission layer blocks it). Parse with the standard extractor:

jq -r '.output[] | select(.type == "message") | .content[] | select(.type == "output_text") | .text' /tmp/xai.json

HTTP=200 with a non-empty body -> use it (SOURCE_PATH=api).

Path C - WebSearch / WebFetch (last resort). Only when both A and B fail (or no key is set at all). Lowest quality (WebSearch favours old high-engagement tweets; WebFetch reads a public profile page). Never reach for it while a key works.

Reason codes (record the true one on every fallback - never write "XAI_API_KEY unavailable" / "sandbox" / "expansion blocked" when a key was set): key-unset (the relevant key was actually empty), http-<code> (non-2xx), empty (200 but no tweets parsed), timeout (curl exceeded --max-time).

Environment Variables

  • TWITTER_API_KEY - twitterapi.io API key (sent as the X-API-Key header). Declared in requires:, so it is injected into this skill's environment; it is the primary fetch path for every branch except list (twitterapi has no list operator). Fast, structured ground-truth with exact engagement counts and real permalinks.
  • XAI_API_KEY - X.AI API key for Grok's x_search tool (declared XAI_API_KEY?, optional). The fallback path when TWITTER_API_KEY is unset or twitterapi returns non-2xx / empty / timeout, and the primary path for the list branch. If both keys are unset, branches degrade to WebSearch/WebFetch at lower quality; the account (all) sub-mode instead hard-exits (TWEET_DIGEST_NO_KEY).

04 同仓库更多 Skills