[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6GXKP6ZX0M8M0FM250ZYN59":3,"skill-same:01M6GXKP6ZX0M8M0FM250ZYN59":97},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":28,"publishedAt":48,"linkOnly":49,"markdownContent":50,"markdownBody":51,"frontMatter":52,"refName":61,"commitSha":62,"skillDirectory":63,"forksCount":64,"repositoryHtmlUrl":65,"repositoryHomepageUrl":66,"ownerHtmlUrl":67,"ownerBlogUrl":66,"ownerDescription":68,"updatedAt":69,"install":70,"resources":82},"01M6GXKP6ZX0M8M0FM250ZYN59","dashboard",null,"Real-time harness observability dashboard. Reads campaigns, fleet sessions, telemetry, and pending queues to present a snapshot of harness state at a glance. Invoked by \u002Fdashboard, \u002Fdo status, or phrases like \"what's happening\" and \"show activity\".","SethGammon\u002FCitadel","SethGammon","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F5523585?v=4","01M6GXJ90FD2CE9QF5GQERWN06","Seth Gammon","01M6GXJ94VXYGK5KAW14EYE901",931,"MIT","https:\u002F\u002Fgithub.com\u002FSethGammon\u002FCitadel\u002Fblob\u002Fmain\u002Fskills\u002Fdashboard\u002FSKILL.md",[18,23],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000013","分析与可视化","SCENARIO","使用场景",{"categoryNo":24,"categoryName":25,"taxonomyType":26,"taxonomyTypeDesc":27},"01JBSEEDCAT000000000000033","数据分析师","PROFESSION","适用职业",[5,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47],"do","fleet","at","activity","false","present","campaigns","time","me","glance","show","observability dashboard","snapshot","show me","real-time","observability","harness","status","sessions","2026-11-10 10:12:30",false,"---\nname: dashboard\nlicense: MIT\ndescription: >-\n  Real-time harness observability dashboard. Reads campaigns, fleet sessions,\n  telemetry, and pending queues to present a snapshot of harness state at a\n  glance. Invoked by \u002Fdashboard, \u002Fdo status, or phrases like \"what's happening\"\n  and \"show activity\".\nuser-invocable: true\nauto-trigger: false\ntrigger_keywords:\n  - dashboard\n  - what's happening\n  - what's going on\n  - show activity\n  - harness state\n  - show me status\nlast-updated: 2026-03-26\n---\n\n# \u002Fdashboard — Harness Observability Dashboard\n\n## When to Use\n\n- \"What's happening?\" \u002F \"Status?\" \u002F \"What's going on?\"\n- \"Show activity\" \u002F \"Show me the dashboard\"\n- After returning to a project after time away\n- When \u002Fdo routes \"status\", \"dashboard\", \"what's happening\", \"what's going on\", \"show activity\"\n- Directly: `\u002Fdashboard`\n\n## Inputs\n\nNone required. Works with whatever state exists on disk.\n\n## Protocol\n\n### Step 0: RUN DASHBOARD IMPLEMENTATION\n\nRun the local dashboard implementation from the project root:\n\n```bash\nnode scripts\u002Fdashboard.js\n```\n\nIf the package scripts are available, this equivalent command is also valid:\n\n```bash\nnpm run dashboard\n```\n\nThe script is read-only. It renders a user-facing control-plane snapshot from\n`.planning\u002F`, telemetry, hook config, coordination state, worktrees, and cost\ndata. Use the manual collection protocol below only as a fallback if the script\nis missing or fails in the current project.\n\n### Step 1: COLLECT STATE\n\nRead the following sources. Each is optional — if a file or directory doesn't\nexist, treat it as empty. Never crash on missing state.\n\n**Campaigns:**\n- Glob `.planning\u002Fcampaigns\u002F*.md`\n- For each file, read the first 40 lines to extract:\n  - `Status:` field\n  - `Direction:` field (truncate to 60 chars)\n  - Phase progress (search for `Phase N of M` or `## Phase` headings)\n  - Most recent line starting with `- [` from the Decision Log\n- If all phases are complete but status is still active, report\n  `needs-completion` and show:\n  `node scripts\u002Fcampaign.js complete \u003Cslug> --archive`\n- If a campaign is marked completed but still lives in `.planning\u002Fcampaigns\u002F`,\n  report `needs-archive` and show the same archive command.\n- If prior build\u002Fverify phases are complete but the `review-package` Exit\n  Evidence row is still pending, missing, or points at a missing local package,\n  report a repair before campaign completion:\n  `node scripts\u002Fpackage-delivery.js \u003Cslug>`\n\n**Cost Data (two sources, prefer real):**\n\n- Primary: run `node scripts\u002Fsession-tokens.js --today` and `--all` — reads Claude Code's native session JSONL for exact token counts\n- Fallback: read `.planning\u002Ftelemetry\u002Fsession-costs.jsonl`; cost priority `real_cost` > `override_cost` > `estimated_cost`; group by `campaign_slug`, sum cost\u002Fagents\u002Fminutes, compute grand total\n- Live session: read `.planning\u002Ftelemetry\u002Fcost-tracker-state.json` for burn rate\n- Label real data \"(real)\" and estimates \"(est)\"\n\n**Fleet Sessions:**\n- Glob `.planning\u002Ffleet\u002Fsession-*.md`\n- For each file, read the first 30 lines to extract:\n  - `status:` field\n  - `wave:` or wave number\n  - `agents:` or agent count\n\n**Recent Telemetry:**\n- Read last 50 lines of `.planning\u002Ftelemetry\u002Fhook-timing.jsonl` (if it exists)\n- Read last 50 lines of `.planning\u002Ftelemetry\u002Faudit.jsonl` (if it exists)\n- Merge and sort by timestamp (descending). Take the 10 most recent entries.\n- For each entry: extract `ts` (or `timestamp`), `hook` (or `event`), and a\n  short description field. Format as relative time.\n\n**Recent Hook Activity (separate from general telemetry):**\n- Read last 20 lines of `.planning\u002Ftelemetry\u002Fhook-timing.jsonl`\n- For `event: \"timing\"` entries: extract `hook`, `duration_ms`, `timestamp` (relative), and `outcome` (pass if no matching error in hook-errors.jsonl within 1s; block if a block entry exists)\n- For `event: \"counter\"` entries: extract metric name as the \"event\" column with count context\n\n**Hook Overhead (timing percentiles):**\n- Read all of `.planning\u002Ftelemetry\u002Fhook-timing.jsonl` (if it exists)\n- Keep only entries with a numeric `duration_ms`; group by `hook`\n- Per hook compute: count, p50, p95, max (nearest-rank percentile over the sorted durations)\n- Sort rows by p95 descending\n- If the file is missing or contains no timed entries, render the one-line note instead\n\n**Routine Quota (account-wide 15 runs \u002F 24h cap):**\n- Read `.planning\u002Ftelemetry\u002Froutine-runs.jsonl` (if it exists)\n- Expected JSONL shape, one record per quota-consuming run:\n  `{\"ts\": \"\u003CISO timestamp>\", \"kind\": \"RemoteTrigger\" | \"CronCreate\" | \"ScheduleWakeup\"}`\n- Count records with `ts` inside the last 24 hours; compare against the cap of 15\n- Warn when the count exceeds 12 (hitting the cap pauses every routine on the\n  account; see `docs\u002FROUTINE-QUOTA.md`)\n- The harness does not write this file automatically yet — remote-run logging\n  populates it when a routine mechanism is actually used. Local runners\n  (`local-watch.js`, `local-daemon.js`, `local-schedule.js`) never consume quota\n  and must not be counted.\n\n**Pending Queues:**\n- Count actionable entries in `.planning\u002Ftelemetry\u002Fdoc-sync-queue.jsonl` where `status` is `pending` or `needs-review` (or 0 if missing)\n- Count lines in `.planning\u002Ftelemetry\u002Fmerge-check-queue.jsonl` (or 0 if missing)\n- Count files in `.planning\u002Fintake\u002F` (or 0 if missing)\n\n**Hook Value Data (for HOOKS VALUE section):**\n- Read `.planning\u002Ftelemetry\u002Fhook-errors.jsonl` (if it exists, last 200 lines)\n  - Count entries where `hook` = \"protect-files\" (blocked file access)\n  - Count entries where `hook` = \"external-action-gate\" (gated external actions)\n  - Count entries where `hook` = \"quality-gate\" (quality violations)\n- Read `.planning\u002Ftelemetry\u002Fhook-timing.jsonl` (if it exists, last 200 lines)\n  - Count entries where `hook` = \"circuit-breaker\" and `metric` = \"trips\"\n  - Count total entries from today (entries containing today's ISO date prefix)\n- Read `.planning\u002Ftelemetry\u002Faudit.jsonl` (if it exists, last 200 lines)\n  - Count entries mentioning \"circuit-breaker\" or \"circuit_breaker\"\n\n**Hook Problem Taxonomy:**\n- Read last 100 entries from `.planning\u002Ftelemetry\u002Fhook-errors.jsonl`.\n- Classify `protect-files` blocks and hard `external-action-gate` blocks as\n  `safety-block` with `info` severity; they prove protection fired and do not\n  create a repair action by themselves.\n- Classify `error` and `parse-fail` actions as `hook-failure` with `high`\n  severity; these are actionable.\n- Classify `blocked-restricted` as `restricted-scope-block` with `high`\n  severity; this is actionable.\n- Classify `first-encounter` and `consent-block` from `external-action-gate` as\n  `approval-needed` with `medium` severity; this is actionable.\n- If an `external-action-gate` approval entry has a later matching\n  `tool-call` entry in `audit.jsonl`, classify it as `resolved-approval` with\n  `info` severity; it should not create a repair action. Treat `git push -u`\n  and `git push` as equivalent for the same branch, and allow a small\n  near-simultaneous timestamp skew between hook and tool-call entries.\n- If an unresolved external approval entry is older than 15 minutes, classify\n  it as `stale-approval` with `low` severity; it should not create a current\n  repair action.\n- Classify entries older than 24 hours as `stale` with `low` severity and do\n  not create a repair action from stale entries.\n- The `\u002Ftelemetry` repair action should appear only when actionable entries are\n  present. Safety blocks remain visible in PROBLEMS and HOOKS VALUE.\n\n**Health:**\n- Count circuit breaker entries from audit.jsonl (from hook value data above)\n- Count total lines in `.planning\u002Ftelemetry\u002Faudit.jsonl` written today\n- Count entries in `hooks` array of `.claude\u002Fhooks-template.json` (or\n  `.claude\u002Fhooks.json` if template not present); use 0 if neither exists\n- Read `.claude\u002Fharness.json` (Citadel's shared config path for all runtimes, including Codex) → `trust` object:\n  - `sessions_completed`, `campaigns_completed` counters\n  - Compute level: novice (sessions \u003C 5), familiar (5-19), trusted (20+ with 2+ campaigns)\n  - If `trust.override` is set, use that and note \"(override)\"\n\n### Step 2: FORMAT RELATIVE TIMESTAMPS\n\nConvert ISO timestamps: \u003C60s → \"just now\" | \u003C60min → \"{N} min ago\" | \u003C24h → \"{N} hr ago\" | else → \"{N} days ago\". Display unparseable timestamps as-is.\n\n### Step 3: RENDER DASHBOARD\n\nOutput verbatim, substituting real values. Always show section headers even when content is \"(none active)\".\n\n```\n=== Citadel Dashboard ===\nAs of: {relative timestamp of most recent event, or \"now\"}\n\nNEXT ACTION\n  Command: {exact command}\n  Why: {why this is next}\n  Confidence: {low | medium | high}\n  Repair available: {yes | no}\n  Runbook: {docs or skill path}\n\nREPAIR CONSOLE\n  {repair|review} | {confidence} | {label}\n    command: {exact command}\n    why: {short reason}\n    runbook: {docs or skill path}\n\nCAMPAIGNS\n  {slug}: Phase {N}\u002F{total} — {direction, max 60 chars, ellipsis if truncated}\n  Last event: {most recent telemetry entry for this campaign, or \"no telemetry\"}\n  (none active)\n\nCOSTS\n  This session: ${cost} | {duration} min | ${rate}\u002Fmin | {messages} msgs | {agents} agents\n  Today:        ${today_total} across {today_sessions} sessions\n  All time:     ${all_time_total} across {all_time_sessions} sessions ({data_source})\n\n  By campaign:\n    {slug}: ${total_cost} across {sessions} sessions ({agents} agents, {minutes} min)\n    _unattached: ${total_cost} across {sessions} sessions\n  (no cost data recorded yet)\n\nROUTINE QUOTA\n  Runs (last 24h): {N}\u002F15\n  WARNING: {N} of 15 routine runs used in the last 24h. Hitting the cap pauses every routine on the account. See docs\u002FROUTINE-QUOTA.md.\n  (remote-run logging populates .planning\u002Ftelemetry\u002Froutine-runs.jsonl - local runners do not consume quota)\n\nHOOKS VALUE\n  Circuit breaker: {N} trips (prevented token spirals)\n  Quality gate:    {N} violations caught pre-commit\n  Protect-files:   {N} blocks (path traversal, secrets)\n  External gate:   {N} actions gated\n  Total hook fires today: {N}\n  (raw facts only -- no inflated savings claims)\n\nFLEET SESSIONS\n  {slug}: Wave {N} — {agent count} agents — {status}\n  (none active)\n\nRECENT ACTIVITY (last 10 events)\n  {relative time} | {hook\u002Fevent name} | {description}\n  (no telemetry recorded yet)\n\nHOOK ACTIVITY (last 10 hook fires)\n  {relative time} | {hook name} | {duration_ms}ms | {outcome: pass\u002Fblock\u002Fwarn}\n  (no hook timing recorded yet — set CITADEL_DEBUG=true in settings.json for verbose output)\n\nHOOK OVERHEAD (sorted by p95 descending)\n  hook                        count      p50      p95      max\n  {hook name}                   {N}   {N}ms    {N}ms    {N}ms\n  (no hook timing data recorded yet)\n\nPROBLEMS\n  Actionable: {N} | Safety blocks: {N} | Resolved approvals: {N} | Stale: {N}\n  {relative time} | {severity} | {category} | {hook name} | {description}\n  (none recorded)\n\nPENDING\n  Doc sync:     {N} items queued\n  Merge reviews: {N} items queued\n  Intake items:  {N} in .planning\u002Fintake\u002F\n\nHEALTH\n  Circuit breaker trips this session: {N}\n  Audit entries today:                {N}\n  Hooks installed:                    {N}\n  Operator tier:                      {novice | familiar | trusted} ({N} sessions, {N} campaigns)\n\nQUICK COMMANDS\n  \u002Fdo continue    — resume active campaign\n  \u002Fdo rollback    — restore last checkpoint\n  \u002Ftelemetry      — cost breakdown, hook activity, telemetry settings\n  \u002Ftriage prs     — review open PRs\n  \u002Fpr-watch       — watch PR CI\n  \u002Flearn          — extract patterns from last completed campaign\n```\n\n### Step 4: FRINGE CASE HANDLING\n\n**`.planning\u002F` missing:** All zeros, \"(none active)\"; add \"Run \u002Fdo setup --express to initialize.\"\n**harness.json missing or malformed:** Show \"not configured\" for hooks count; do not crash.\n**Malformed campaign file:** Skip it; note `(N campaign file(s) skipped — malformed)`.\n**Large telemetry files:** Read last 50 lines only.\n**Missing timestamps:** Fall back to file modification time; display entry without timestamp if unavailable.\n**All campaigns completed:** Note \"No active campaigns\" at top of CAMPAIGNS section.\n**Completed campaign still active:** Show the exact `node scripts\u002Fcampaign.js complete \u003Cslug> --archive` repair command; suggesting `\u002Fdo continue` here is wrong because the campaign is already finished.\n**Campaign ready for review package:** Show the exact `node scripts\u002Fpackage-delivery.js \u003Cslug>` repair command before showing campaign completion.\n**All fleet sessions idle:** Note \"No active fleet sessions\" under FLEET SESSIONS.\n**routine-runs.jsonl missing or no runs in window:** Show `Runs (last 24h): 0\u002F15` plus the one-line population hint; only show the WARNING line when more than 12 runs are counted.\n**Mixed state:** Proceed with whatever state exists; note each missing directory inline.\n**Doc-sync backlog:** Surface `\u002Flearn --doc-sync` as a repair action with `skills\u002Flearn\u002FSKILL.md` as runbook.\n**Dirty worktree:** Surface `git status --short` as a review action; do not suggest destructive cleanup.\n**Only safety blocks recorded:** Show them in PROBLEMS and HOOKS VALUE, but do not surface `\u002Ftelemetry` as NEXT ACTION.\n**Actionable hook problem recorded:** Surface `\u002Ftelemetry` as repair action with `skills\u002Ftelemetry\u002FSKILL.md` as runbook.\n\n## Contextual Gates\n\n**Disclosure:** \"Displaying harness dashboard. No files modified.\"\n**Reversibility:** green — read-only; no files modified\n**Trust gates:**\n- Any: view the full dashboard\n\n## Quality Gates\n\n- Dashboard must render even when all state files are missing\n- Never display raw JSON to the user — always parse and format\n- Relative timestamps required — never show raw ISO strings in output\n- Campaign direction truncated to 60 chars with \"...\" if longer\n- NEXT ACTION must include command, why, confidence, repair availability, and runbook when known\n- REPAIR CONSOLE must list actionable repairs before raw activity logs\n- Safety blocks must not be treated as urgent repairs unless paired with an actionable hook failure, approval, or restricted-scope block\n- Total output must be skimmable in under 30 seconds\n\n## Exit Protocol\n\n\u002Fdashboard does not produce a HANDOFF block. It is a read-only observability\ntool. After displaying the dashboard, wait for the next user command.\n","# \u002Fdashboard — Harness Observability Dashboard\n\n## When to Use\n\n- \"What's happening?\" \u002F \"Status?\" \u002F \"What's going on?\"\n- \"Show activity\" \u002F \"Show me the dashboard\"\n- After returning to a project after time away\n- When \u002Fdo routes \"status\", \"dashboard\", \"what's happening\", \"what's going on\", \"show activity\"\n- Directly: `\u002Fdashboard`\n\n## Inputs\n\nNone required. Works with whatever state exists on disk.\n\n## Protocol\n\n### Step 0: RUN DASHBOARD IMPLEMENTATION\n\nRun the local dashboard implementation from the project root:\n\n```bash\nnode scripts\u002Fdashboard.js\n```\n\nIf the package scripts are available, this equivalent command is also valid:\n\n```bash\nnpm run dashboard\n```\n\nThe script is read-only. It renders a user-facing control-plane snapshot from\n`.planning\u002F`, telemetry, hook config, coordination state, worktrees, and cost\ndata. Use the manual collection protocol below only as a fallback if the script\nis missing or fails in the current project.\n\n### Step 1: COLLECT STATE\n\nRead the following sources. Each is optional — if a file or directory doesn't\nexist, treat it as empty. Never crash on missing state.\n\n**Campaigns:**\n- Glob `.planning\u002Fcampaigns\u002F*.md`\n- For each file, read the first 40 lines to extract:\n  - `Status:` field\n  - `Direction:` field (truncate to 60 chars)\n  - Phase progress (search for `Phase N of M` or `## Phase` headings)\n  - Most recent line starting with `- [` from the Decision Log\n- If all phases are complete but status is still active, report\n  `needs-completion` and show:\n  `node scripts\u002Fcampaign.js complete \u003Cslug> --archive`\n- If a campaign is marked completed but still lives in `.planning\u002Fcampaigns\u002F`,\n  report `needs-archive` and show the same archive command.\n- If prior build\u002Fverify phases are complete but the `review-package` Exit\n  Evidence row is still pending, missing, or points at a missing local package,\n  report a repair before campaign completion:\n  `node scripts\u002Fpackage-delivery.js \u003Cslug>`\n\n**Cost Data (two sources, prefer real):**\n\n- Primary: run `node scripts\u002Fsession-tokens.js --today` and `--all` — reads Claude Code's native session JSONL for exact token counts\n- Fallback: read `.planning\u002Ftelemetry\u002Fsession-costs.jsonl`; cost priority `real_cost` > `override_cost` > `estimated_cost`; group by `campaign_slug`, sum cost\u002Fagents\u002Fminutes, compute grand total\n- Live session: read `.planning\u002Ftelemetry\u002Fcost-tracker-state.json` for burn rate\n- Label real data \"(real)\" and estimates \"(est)\"\n\n**Fleet Sessions:**\n- Glob `.planning\u002Ffleet\u002Fsession-*.md`\n- For each file, read the first 30 lines to extract:\n  - `status:` field\n  - `wave:` or wave number\n  - `agents:` or agent count\n\n**Recent Telemetry:**\n- Read last 50 lines of `.planning\u002Ftelemetry\u002Fhook-timing.jsonl` (if it exists)\n- Read last 50 lines of `.planning\u002Ftelemetry\u002Faudit.jsonl` (if it exists)\n- Merge and sort by timestamp (descending). Take the 10 most recent entries.\n- For each entry: extract `ts` (or `timestamp`), `hook` (or `event`), and a\n  short description field. Format as relative time.\n\n**Recent Hook Activity (separate from general telemetry):**\n- Read last 20 lines of `.planning\u002Ftelemetry\u002Fhook-timing.jsonl`\n- For `event: \"timing\"` entries: extract `hook`, `duration_ms`, `timestamp` (relative), and `outcome` (pass if no matching error in hook-errors.jsonl within 1s; block if a block entry exists)\n- For `event: \"counter\"` entries: extract metric name as the \"event\" column with count context\n\n**Hook Overhead (timing percentiles):**\n- Read all of `.planning\u002Ftelemetry\u002Fhook-timing.jsonl` (if it exists)\n- Keep only entries with a numeric `duration_ms`; group by `hook`\n- Per hook compute: count, p50, p95, max (nearest-rank percentile over the sorted durations)\n- Sort rows by p95 descending\n- If the file is missing or contains no timed entries, render the one-line note instead\n\n**Routine Quota (account-wide 15 runs \u002F 24h cap):**\n- Read `.planning\u002Ftelemetry\u002Froutine-runs.jsonl` (if it exists)\n- Expected JSONL shape, one record per quota-consuming run:\n  `{\"ts\": \"\u003CISO timestamp>\", \"kind\": \"RemoteTrigger\" | \"CronCreate\" | \"ScheduleWakeup\"}`\n- Count records with `ts` inside the last 24 hours; compare against the cap of 15\n- Warn when the count exceeds 12 (hitting the cap pauses every routine on the\n  account; see `docs\u002FROUTINE-QUOTA.md`)\n- The harness does not write this file automatically yet — remote-run logging\n  populates it when a routine mechanism is actually used. Local runners\n  (`local-watch.js`, `local-daemon.js`, `local-schedule.js`) never consume quota\n  and must not be counted.\n\n**Pending Queues:**\n- Count actionable entries in `.planning\u002Ftelemetry\u002Fdoc-sync-queue.jsonl` where `status` is `pending` or `needs-review` (or 0 if missing)\n- Count lines in `.planning\u002Ftelemetry\u002Fmerge-check-queue.jsonl` (or 0 if missing)\n- Count files in `.planning\u002Fintake\u002F` (or 0 if missing)\n\n**Hook Value Data (for HOOKS VALUE section):**\n- Read `.planning\u002Ftelemetry\u002Fhook-errors.jsonl` (if it exists, last 200 lines)\n  - Count entries where `hook` = \"protect-files\" (blocked file access)\n  - Count entries where `hook` = \"external-action-gate\" (gated external actions)\n  - Count entries where `hook` = \"quality-gate\" (quality violations)\n- Read `.planning\u002Ftelemetry\u002Fhook-timing.jsonl` (if it exists, last 200 lines)\n  - Count entries where `hook` = \"circuit-breaker\" and `metric` = \"trips\"\n  - Count total entries from today (entries containing today's ISO date prefix)\n- Read `.planning\u002Ftelemetry\u002Faudit.jsonl` (if it exists, last 200 lines)\n  - Count entries mentioning \"circuit-breaker\" or \"circuit_breaker\"\n\n**Hook Problem Taxonomy:**\n- Read last 100 entries from `.planning\u002Ftelemetry\u002Fhook-errors.jsonl`.\n- Classify `protect-files` blocks and hard `external-action-gate` blocks as\n  `safety-block` with `info` severity; they prove protection fired and do not\n  create a repair action by themselves.\n- Classify `error` and `parse-fail` actions as `hook-failure` with `high`\n  severity; these are actionable.\n- Classify `blocked-restricted` as `restricted-scope-block` with `high`\n  severity; this is actionable.\n- Classify `first-encounter` and `consent-block` from `external-action-gate` as\n  `approval-needed` with `medium` severity; this is actionable.\n- If an `external-action-gate` approval entry has a later matching\n  `tool-call` entry in `audit.jsonl`, classify it as `resolved-approval` with\n  `info` severity; it should not create a repair action. Treat `git push -u`\n  and `git push` as equivalent for the same branch, and allow a small\n  near-simultaneous timestamp skew between hook and tool-call entries.\n- If an unresolved external approval entry is older than 15 minutes, classify\n  it as `stale-approval` with `low` severity; it should not create a current\n  repair action.\n- Classify entries older than 24 hours as `stale` with `low` severity and do\n  not create a repair action from stale entries.\n- The `\u002Ftelemetry` repair action should appear only when actionable entries are\n  present. Safety blocks remain visible in PROBLEMS and HOOKS VALUE.\n\n**Health:**\n- Count circuit breaker entries from audit.jsonl (from hook value data above)\n- Count total lines in `.planning\u002Ftelemetry\u002Faudit.jsonl` written today\n- Count entries in `hooks` array of `.claude\u002Fhooks-template.json` (or\n  `.claude\u002Fhooks.json` if template not present); use 0 if neither exists\n- Read `.claude\u002Fharness.json` (Citadel's shared config path for all runtimes, including Codex) → `trust` object:\n  - `sessions_completed`, `campaigns_completed` counters\n  - Compute level: novice (sessions \u003C 5), familiar (5-19), trusted (20+ with 2+ campaigns)\n  - If `trust.override` is set, use that and note \"(override)\"\n\n### Step 2: FORMAT RELATIVE TIMESTAMPS\n\nConvert ISO timestamps: \u003C60s → \"just now\" | \u003C60min → \"{N} min ago\" | \u003C24h → \"{N} hr ago\" | else → \"{N} days ago\". Display unparseable timestamps as-is.\n\n### Step 3: RENDER DASHBOARD\n\nOutput verbatim, substituting real values. Always show section headers even when content is \"(none active)\".\n\n```\n=== Citadel Dashboard ===\nAs of: {relative timestamp of most recent event, or \"now\"}\n\nNEXT ACTION\n  Command: {exact command}\n  Why: {why this is next}\n  Confidence: {low | medium | high}\n  Repair available: {yes | no}\n  Runbook: {docs or skill path}\n\nREPAIR CONSOLE\n  {repair|review} | {confidence} | {label}\n    command: {exact command}\n    why: {short reason}\n    runbook: {docs or skill path}\n\nCAMPAIGNS\n  {slug}: Phase {N}\u002F{total} — {direction, max 60 chars, ellipsis if truncated}\n  Last event: {most recent telemetry entry for this campaign, or \"no telemetry\"}\n  (none active)\n\nCOSTS\n  This session: ${cost} | {duration} min | ${rate}\u002Fmin | {messages} msgs | {agents} agents\n  Today:        ${today_total} across {today_sessions} sessions\n  All time:     ${all_time_total} across {all_time_sessions} sessions ({data_source})\n\n  By campaign:\n    {slug}: ${total_cost} across {sessions} sessions ({agents} agents, {minutes} min)\n    _unattached: ${total_cost} across {sessions} sessions\n  (no cost data recorded yet)\n\nROUTINE QUOTA\n  Runs (last 24h): {N}\u002F15\n  WARNING: {N} of 15 routine runs used in the last 24h. Hitting the cap pauses every routine on the account. See docs\u002FROUTINE-QUOTA.md.\n  (remote-run logging populates .planning\u002Ftelemetry\u002Froutine-runs.jsonl - local runners do not consume quota)\n\nHOOKS VALUE\n  Circuit breaker: {N} trips (prevented token spirals)\n  Quality gate:    {N} violations caught pre-commit\n  Protect-files:   {N} blocks (path traversal, secrets)\n  External gate:   {N} actions gated\n  Total hook fires today: {N}\n  (raw facts only -- no inflated savings claims)\n\nFLEET SESSIONS\n  {slug}: Wave {N} — {agent count} agents — {status}\n  (none active)\n\nRECENT ACTIVITY (last 10 events)\n  {relative time} | {hook\u002Fevent name} | {description}\n  (no telemetry recorded yet)\n\nHOOK ACTIVITY (last 10 hook fires)\n  {relative time} | {hook name} | {duration_ms}ms | {outcome: pass\u002Fblock\u002Fwarn}\n  (no hook timing recorded yet — set CITADEL_DEBUG=true in settings.json for verbose output)\n\nHOOK OVERHEAD (sorted by p95 descending)\n  hook                        count      p50      p95      max\n  {hook name}                   {N}   {N}ms    {N}ms    {N}ms\n  (no hook timing data recorded yet)\n\nPROBLEMS\n  Actionable: {N} | Safety blocks: {N} | Resolved approvals: {N} | Stale: {N}\n  {relative time} | {severity} | {category} | {hook name} | {description}\n  (none recorded)\n\nPENDING\n  Doc sync:     {N} items queued\n  Merge reviews: {N} items queued\n  Intake items:  {N} in .planning\u002Fintake\u002F\n\nHEALTH\n  Circuit breaker trips this session: {N}\n  Audit entries today:                {N}\n  Hooks installed:                    {N}\n  Operator tier:                      {novice | familiar | trusted} ({N} sessions, {N} campaigns)\n\nQUICK COMMANDS\n  \u002Fdo continue    — resume active campaign\n  \u002Fdo rollback    — restore last checkpoint\n  \u002Ftelemetry      — cost breakdown, hook activity, telemetry settings\n  \u002Ftriage prs     — review open PRs\n  \u002Fpr-watch       — watch PR CI\n  \u002Flearn          — extract patterns from last completed campaign\n```\n\n### Step 4: FRINGE CASE HANDLING\n\n**`.planning\u002F` missing:** All zeros, \"(none active)\"; add \"Run \u002Fdo setup --express to initialize.\"\n**harness.json missing or malformed:** Show \"not configured\" for hooks count; do not crash.\n**Malformed campaign file:** Skip it; note `(N campaign file(s) skipped — malformed)`.\n**Large telemetry files:** Read last 50 lines only.\n**Missing timestamps:** Fall back to file modification time; display entry without timestamp if unavailable.\n**All campaigns completed:** Note \"No active campaigns\" at top of CAMPAIGNS section.\n**Completed campaign still active:** Show the exact `node scripts\u002Fcampaign.js complete \u003Cslug> --archive` repair command; suggesting `\u002Fdo continue` here is wrong because the campaign is already finished.\n**Campaign ready for review package:** Show the exact `node scripts\u002Fpackage-delivery.js \u003Cslug>` repair command before showing campaign completion.\n**All fleet sessions idle:** Note \"No active fleet sessions\" under FLEET SESSIONS.\n**routine-runs.jsonl missing or no runs in window:** Show `Runs (last 24h): 0\u002F15` plus the one-line population hint; only show the WARNING line when more than 12 runs are counted.\n**Mixed state:** Proceed with whatever state exists; note each missing directory inline.\n**Doc-sync backlog:** Surface `\u002Flearn --doc-sync` as a repair action with `skills\u002Flearn\u002FSKILL.md` as runbook.\n**Dirty worktree:** Surface `git status --short` as a review action; do not suggest destructive cleanup.\n**Only safety blocks recorded:** Show them in PROBLEMS and HOOKS VALUE, but do not surface `\u002Ftelemetry` as NEXT ACTION.\n**Actionable hook problem recorded:** Surface `\u002Ftelemetry` as repair action with `skills\u002Ftelemetry\u002FSKILL.md` as runbook.\n\n## Contextual Gates\n\n**Disclosure:** \"Displaying harness dashboard. No files modified.\"\n**Reversibility:** green — read-only; no files modified\n**Trust gates:**\n- Any: view the full dashboard\n\n## Quality Gates\n\n- Dashboard must render even when all state files are missing\n- Never display raw JSON to the user — always parse and format\n- Relative timestamps required — never show raw ISO strings in output\n- Campaign direction truncated to 60 chars with \"...\" if longer\n- NEXT ACTION must include command, why, confidence, repair availability, and runbook when known\n- REPAIR CONSOLE must list actionable repairs before raw activity logs\n- Safety blocks must not be treated as urgent repairs unless paired with an actionable hook failure, approval, or restricted-scope block\n- Total output must be skimmable in under 30 seconds\n\n## Exit Protocol\n\n\u002Fdashboard does not produce a HANDOFF block. It is a read-only observability\ntool. After displaying the dashboard, wait for the next user command.\n",{"name":5,"license":15,"description":7,"auto-trigger":49,"last-updated":53,"user-invocable":54,"trigger_keywords":55},1774483200000,true,[5,56,57,58,59,60],"what's happening","what's going on","show activity","harness state","show me status","main","e1a3d45697a0f323aaedcbea0ab6a0387b5fe722","skills\u002Fdashboard",84,"https:\u002F\u002Fgithub.com\u002FSethGammon\u002FCitadel","https:\u002F\u002Fsethgammon.github.io\u002FCitadel\u002F","https:\u002F\u002Fgithub.com\u002FSethGammon","Building Citadel, the repo-local operating layer for Claude Code and Codex. Persistent work, routing, verification, and coordinated agent fleets.","2026-11-13 02:05:13",{"source":8,"skillName":5,"commands":71,"prompt":81,"zipAvailable":54},[72,75,78],{"tool":73,"command":74},"npx","npx skills add SethGammon\u002FCitadel --skill dashboard",{"tool":76,"command":77},"bunx","bunx skills add SethGammon\u002FCitadel --skill dashboard",{"tool":79,"command":80},"pnpm","pnpm dlx skills add SethGammon\u002FCitadel --skill dashboard","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FSethGammon\u002FCitadel\u002Fblob\u002Fmain\u002Fskills\u002Fdashboard\u002FSKILL.md",[83,89,93],{"path":84,"resourceType":85,"resourceTypeDesc":86,"sizeBytes":87,"sourceHtmlUrl":88},"__benchmarks__\u002Factive-campaign.md","REFERENCE","参考文档",833,"https:\u002F\u002Fgithub.com\u002FSethGammon\u002FCitadel\u002Fblob\u002Fmain\u002Fskills\u002Fdashboard\u002F__benchmarks__\u002Factive-campaign.md",{"path":90,"resourceType":85,"resourceTypeDesc":86,"sizeBytes":91,"sourceHtmlUrl":92},"__benchmarks__\u002Fno-planning-dir.md",977,"https:\u002F\u002Fgithub.com\u002FSethGammon\u002FCitadel\u002Fblob\u002Fmain\u002Fskills\u002Fdashboard\u002F__benchmarks__\u002Fno-planning-dir.md",{"path":94,"resourceType":85,"resourceTypeDesc":86,"sizeBytes":95,"sourceHtmlUrl":96},"__benchmarks__\u002Fvague-status-phrase.md",817,"https:\u002F\u002Fgithub.com\u002FSethGammon\u002FCitadel\u002Fblob\u002Fmain\u002Fskills\u002Fdashboard\u002F__benchmarks__\u002Fvague-status-phrase.md",[98,121,148,171,197,216],{"skillNo":99,"name":100,"nameZh":6,"description":101,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":102,"categories":103,"keywords":104,"publishedAt":48},"01M6GXSBKE9A2HQWF8QRAZ4WJ5","workspace","Multi-repo campaign coordinator. Same lifecycle as fleet -- scope claims, discovery relay, wave-based execution -- but the unit of work is a repo, not a file. Coordinates campaigns across repositories with shared context.","https:\u002F\u002Fgithub.com\u002FSethGammon\u002FCitadel\u002Fblob\u002Fmain\u002Fskills\u002Fworkspace\u002FSKILL.md",[],[100,105,106,107,108,109,110,111,112,113,35,114,115,33,116,117,118,119,30,120],"repos","snowflake","multi-repo","work","context","campaign","coordinate","but","across","shared","as","scope","add","unit","coordinator","claims",{"skillNo":122,"name":123,"nameZh":6,"description":124,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":125,"categories":126,"keywords":130,"publishedAt":48},"01M6GXS873AD1Y6RTM6GV47F3E","wiki","Markdown-first knowledge base where the LLM acts as librarian. Ingests raw sources, compiles and interlinks topic files, self-maintains an index. No vector DB or embeddings required -- uses LLM-native navigation over structured markdown up to ~400K words.","https:\u002F\u002Fgithub.com\u002FSethGammon\u002FCitadel\u002Fblob\u002Fmain\u002Fskills\u002Fwiki\u002FSKILL.md",[127],{"categoryNo":128,"categoryName":129,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000021","资料检索与研究",[123,131,132,133,134,135,136,137,138,139,140,141,142,143,144,145,115,33,146,147],"vector","embeddings","librarian","markdown","knowledge","karpathy","knowledge management","karpathy wiki","up","files","maintain","self","knowledge base","project","db","sources","native",{"skillNo":149,"name":150,"nameZh":6,"description":151,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":152,"categories":153,"keywords":154,"publishedAt":48},"01M6GXS2840SJ61285SZ43EJC3","watch","File sentinel that monitors the working directory for changes and marker comments, then auto-triggers appropriate skills. Poll-based via git diff against the last scan commit. Writes intake items for batch processing and routes marker actions through \u002Fdo. Use for automatic reactions to file changes; do NOT use for one-off inspection or tasks needing human judgment per file.","https:\u002F\u002Fgithub.com\u002FSethGammon\u002FCitadel\u002Fblob\u002Fmain\u002Fskills\u002Fwatch\u002FSKILL.md",[],[150,155,156,157,158,159,160,161,162,163,164,165,140,29,166,167,168,169,33,170],"directory","commit","tasks","git","batch","actions","start","sentinel","diff","auto","stop","scan","human","judgment","reactions","processing",{"skillNo":172,"name":173,"nameZh":6,"description":174,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":175,"categories":176,"keywords":180,"publishedAt":48},"01M6GXRXHYJ5RD9RZP81S2VFSW","verify","Self-test the Citadel hook pipeline from within a live session. Exercises real tool calls (Write, Edit, Bash, Read) and checks that hooks fired, telemetry accumulated, and no errors occurred. Reports HOOK HEALTH: PASS or HOOK HEALTH: FAIL with a per-hook breakdown.","https:\u002F\u002Fgithub.com\u002FSethGammon\u002FCitadel\u002Fblob\u002Fmain\u002Fskills\u002Fverify\u002FSKILL.md",[177],{"categoryNo":178,"categoryName":179,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000008","测试与质量",[173,181,182,45,183,184,185,186,187,188,189,190,191,142,192,33,193,194,195,196],"citadel","telemetry","pipeline","Bash","tool","read","write","edit","session","hooks","reports","live","hook","errors","calls","true",{"skillNo":198,"name":199,"nameZh":6,"description":200,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":201,"categories":202,"keywords":203,"publishedAt":48},"01M6GXRS9MCJFMQZ917H2XH3AC","unharness","Safely leave Citadel using the active adoption receipt. Produces a no-write, reviewable plan, preserves a portable archive, removes only exact owned material, and reports modified or externally registered surfaces as retained or unknown. Legacy installs must be imported before exact leave is claimed.","https:\u002F\u002Fgithub.com\u002FSethGammon\u002FCitadel\u002Fblob\u002Fmain\u002Fskills\u002Funharness\u002FSKILL.md",[],[199,204,181,45,205,206,207,208,209,139,191,210,211,212,213,214,115,33,215,196],"uninstall","adoption","plan","archive","remove","only","Legacy","surfaces","clean","imported","active","material",{"skillNo":217,"name":218,"nameZh":6,"description":219,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":220,"categories":221,"keywords":228,"publishedAt":48},"01M6GXRMEJ2VHD43R9FNPD4910","triage","GitHub issue and PR investigator. Pulls open issues\u002FPRs, classifies them, searches the codebase for root cause or reviews contributed code, proposes fixes with file:line references, and optionally implements fixes. Use for investigating GitHub issues and reviewing PRs; do NOT use for general code review unrelated to GitHub issues.","https:\u002F\u002Fgithub.com\u002FSethGammon\u002FCitadel\u002Fblob\u002Fmain\u002Fskills\u002Ftriage\u002FSKILL.md",[222,225],{"categoryNo":223,"categoryName":224,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000007","代码审查",{"categoryNo":226,"categoryName":227,"taxonomyType":26,"taxonomyTypeDesc":27},"01JBSEEDCAT000000000000030","开发工程师",[218,229,230,231,232,233,234,235,236,237,238,239,240,241,29,33,242,243,244,196],"code review","review","review pr","pr","code","github issues","issues","open","use","unrelated","issue","general","references","line","investigator","reviews"]