[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6PAN48MA2MGDZT0N7JEV02H":3,"skill-same:01M6PAN48MA2MGDZT0N7JEV02H":93},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":15,"categories":16,"keywords":22,"publishedAt":42,"linkOnly":43,"markdownContent":44,"markdownBody":45,"frontMatter":46,"refName":30,"commitSha":74,"skillDirectory":75,"forksCount":13,"repositoryHtmlUrl":76,"repositoryHomepageUrl":6,"ownerHtmlUrl":77,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":78,"install":79,"resources":92},"01M6PAN48MA2MGDZT0N7JEV02H","nika",null,"Runs repeatable AI work as checked, budgeted workflow files.","duc-duy-1612\u002FOnline-Delivery-Systems-Web","duc-duy-1612","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F192231384?v=4","01M6NSDKQQDKVTHZ8DZMZ1VYY6","01M6NSDKW8DDVTQHBT1HBDDJPT",0,"MIT","https:\u002F\u002Fgithub.com\u002Fduc-duy-1612\u002FOnline-Delivery-Systems-Web\u002Fblob\u002Fmain\u002F.agents\u002Fskills_archive\u002Fnika\u002FSKILL.md",[17],{"categoryNo":18,"categoryName":19,"taxonomyType":20,"taxonomyTypeDesc":21},"01JBSEEDCAT000000000000023","Skill与智能体开发","SCENARIO","使用场景",[5,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41],"files","tree","studio","critical","as","local","deterministic","main","community","MCP","AI-agents","cost","autonomous","work","license","code","automation","workflow","linux","2026-11-10 10:12:30",false,"---\nname: nika\ndescription: \"Runs repeatable AI work as checked, budgeted workflow files.\"\nrisk: critical\nsource: https:\u002F\u002Fgithub.com\u002Fsupernovae-st\u002Fnika-agents\u002Ftree\u002Fmain\u002Fskills\u002Fnika\ndate_added: \"2026-09-04\"\nsource_repo: supernovae-st\u002Fnika-agents\nsource_type: community\nversion: 1.1.0\nauthor: Thibaut Melen (@ThibautMelen) · SuperNovae Studio (github.com\u002Fsupernovae-st)\nlicense: MIT\nlicense_source: https:\u002F\u002Fgithub.com\u002Fsupernovae-st\u002Fnika-agents\u002Fblob\u002Fmain\u002FLICENSE\nplatforms: [linux, macos]\nprerequisites:\n  commands: [nika]\nmetadata:\n  hermes:\n    tags: [Workflow, Automation, Deterministic, Cost-Control, Audit, Local-First, MCP]\n    category: autonomous-ai-agents\n    related_skills: [opencode, claude-code, codex]\n    homepage: https:\u002F\u002Fnika.sh\n    requires_toolsets: [terminal]\n---\n\n# Nika Skill\n\nUse [Nika](https:\u002F\u002Fnika.sh) as a deterministic workflow worker orchestrated by\nthe Hermes `terminal` tool. Nika is an open-source (AGPL) Rust engine that captures\na repeatable AI task as a plain-text `*.nika.yaml` file, audits it **before a\nsingle token is spent** (plan, cost floor, secret flows, types), executes it\nagainst local or cloud providers (Ollama\u002Fllama.cpp\u002FvLLM included), and records\na tamper-evident trace.\n\nDivision of labor: **Hermes orchestrates · Nika captures repeatable work as a\ncheckable file and runs it with receipts.** Nika is NOT another coding agent —\nfor autonomous coding, use the `opencode` skill. Delegate to Nika when the\nwork should be *repeatable, budgeted, and auditable*.\n\n## When to Use\n\n- The user asks to run, check, or author a `*.nika.yaml` workflow\n- A task will be repeated (daily digest, triage, ETL, report, multi-step LLM\n  pipeline) — capture it as a workflow instead of re-prompting\n- The user wants a hard cost cap, a cost estimate before running, or\n  receipts\u002Faudit of what ran\n- A pipeline mixes models\u002Fproviders (local + cloud) or mixes LLM steps with\n  shell\u002FHTTP\u002Ffile steps\n- The user wants a run they can replay, verify, or reproduce later\n\n### When NOT to use\n\n- One-off questions or single tool calls — just answer or use a tool\n- Autonomous code implementation\u002Frefactoring\u002FPR review — use the `opencode` skill\n- Interactive back-and-forth tasks — workflows are non-interactive by design\n\n## Prerequisites\n\n- Nika installed: `brew install supernovae-st\u002Ftap\u002Fnika` — other install\n  paths (script, manual download) are documented at https:\u002F\u002Fnika.sh: installing\n  is a human step, not something this skill runs\n- Verify: `terminal(command=\"nika --version\")`\n- Zero keys needed for local\u002Foffline work: `--model mock\u002Fecho` (offline) and\n  `--model ollama\u002F...` (local) run without any API key\n- Cloud providers read standard env vars from the shell;\n  `terminal(command=\"nika doctor\")` diagnoses and prints exact fix commands\n\n## How to Run\n\nProve the toolchain offline first (no key, no network):\n\n```\nterminal(command=\"nika examples run 01-hello --model mock\u002Fecho\")\n```\n\nRun a real workflow — local model first:\n\n```\nterminal(command=\"nika run flow.nika.yaml --model ollama\u002Fqwen3.5:4b\", workdir=\"~\u002Fproject\")\n```\n\nCloud model with a hard budget (always set one for paid models):\n\n```\nterminal(command=\"nika run flow.nika.yaml --model mistral\u002Fmistral-small-latest --max-cost-usd 0.25\", workdir=\"~\u002Fproject\")\n```\n\nPass workflow variables:\n\n```\nterminal(command=\"nika run report.nika.yaml --var city=Paris --var days=7 --max-cost-usd 0.50\", workdir=\"~\u002Fproject\")\n```\n\nLong runs: launch in background and poll — do not block the turn:\n\n```\nterminal(command=\"nika run long.nika.yaml --max-cost-usd 1.00\", workdir=\"~\u002Fproject\", background=true)\nprocess(action=\"poll\", session_id=\"\u003Cid>\")\nprocess(action=\"log\", session_id=\"\u003Cid>\")\n```\n\n### The check-before-run law\n\nNever run a workflow you have not checked. `nika check` is a static pre-flight\n(no tokens spent, no network): plan shape, cost floor, secret-flow analysis,\ntype checks, tool args.\n\n```\nterminal(command=\"nika check flow.nika.yaml --json\", workdir=\"~\u002Fproject\")\n```\n\nFindings carry `NIKA-XXXX` codes that explain themselves via\n`nika explain NIKA-XXXX`. Exit 0 = green, safe to run. Fix findings before\nrunning — never suppress them.\n\n### Authoring a workflow\n\nTurn a repeated task into a file. List templates, then instantiate:\n\n```\nterminal(command=\"nika new --from '?'\")\nterminal(command=\"nika new flow.nika.yaml --from chain\", workdir=\"~\u002Fproject\")\n```\n\n`--from` also accepts plain-words intent. Edit the skeleton (`vars:`,\n`tasks:`, `outputs:`), then **check it**. `nika explain flow.nika.yaml`\nnarrates what it will do, the waves, the cost floor, and what it touches —\nbefore anything runs.\n\nThe artifact you are producing looks like this (checks clean on 0.98):\n\n```yaml\nnika: v1\nworkflow: daily-brief\nmodel: ollama\u002Fqwen3.5:4b\ntasks:\n  - id: fetch\n    invoke:\n      tool: \"nika:fetch\"\n      args: { url: \"https:\u002F\u002Fhn.algolia.com\u002Fapi\u002Fv1\u002Fsearch?tags=front_page\" }\n  - id: brief\n    depends_on: [fetch]\n    infer:\n      max_tokens: 300\n      prompt: |\n        Five bullet points, most signal first: ${{ tasks.fetch.output }}\noutputs:\n  brief: ${{ tasks.brief.output }}\n```\n\nOne file, plain YAML: tasks, an explicit dependency, a bounded model step,\na declared output. That file is what gets checked, run, diffed and reused.\n\n### Cost honesty\n\n- When the workflow prices above the budget, `--max-cost-usd` refuses to\n  start (exit 2, zero tokens) — and since 0.99 the pre-start floor prices\n  the EFFECTIVE model, `--model` override included\n- Mid-run, the ledger stops the workflow the moment real spend crosses the\n  budget: the crossing call completes, nothing new starts, the run fails\n  `NIKA-1704` (exit 1) with spent-vs-budget\n- Estimates use LIST RATES from the vendored public catalog; local · mock ·\n  unpriced work is never blocked\n- A model absent from the catalog meters as $0 — a paid *uncataloged* model\n  runs with no budget protection; prefer cataloged ids (`nika catalog`)\n- Report the cost line from the final run card (the summary block `nika\n  run` prints last — status, cost, trace path) back to the user verbatim\n\n### Receipts and verification\n\nEvery run writes a trace under `.nika\u002Ftraces\u002F` — the run card prints the\ntrace path on its `trace:` line. Both commands take that path (bare\ninvocations are a usage error):\n\n```\nterminal(command=\"nika trace show .nika\u002Ftraces\u002F\u003Crun>.ndjson\", workdir=\"~\u002Fproject\")\nterminal(command=\"nika trace verify .nika\u002Ftraces\u002F\u003Crun>.ndjson\", workdir=\"~\u002Fproject\")\n```\n\n`trace verify` checks the tamper-evidence hash chain: exit 0 intact · 2\nbroken · 3 pre-chain. Also useful: `nika trace outputs` · `nika trace flow` ·\n`nika trace reproduce` · `nika trace export` (OTLP lines).\n\n### Optional: MCP oracle tools\n\nNika also ships a read-only MCP oracle (`nika mcp`) exposing validation and\nlearning tools (`nika_check`, `nika_explain`, `nika_schema`, `nika_examples`,\n`nika_template`, `nika_canon`, `nika_catalog`, `nika_tools`). If the user\nwants those wired into their agent client, point them at the wiring guide —\nhttps:\u002F\u002Fgithub.com\u002Fsupernovae-st\u002Fnika-agents\u002Ftree\u002Fmain\u002Fintegrations\u002Fmcp —\nediting the client's own configuration is the user's step, never this\nskill's. Without the oracle, everything above still works over the terminal;\nrunning workflows stays there regardless, where the budget flags and traces\nlive.\n\n## Quick Reference\n\n| Command | Use |\n|---------|-----|\n| `nika welcome` | What Nika is + what this machine has (offline, exit 0) |\n| `nika new \u003Cfile> --from \u003Ctemplate>` | Scaffold a workflow (`--from '?'` lists) |\n| `nika check \u003Cfile> --json` | Static pre-flight — ALWAYS before run |\n| `nika explain \u003Cfile>` | Narrate: waves, cost floor, touches |\n| `nika run \u003Cfile> --model \u003Cp\u002Fm> --max-cost-usd \u003Cusd>` | Execute with budget |\n| `nika test \u003Cfile>` | Golden test under the mock provider (offline) |\n| `nika trace show\u002Fverify\u002Foutputs\u002Fflow \u003Ctrace>` | Receipts after a run (path from the run card's `trace:` line) |\n| `nika doctor` | Diagnose env\u002Fkeys — prints exact fixes |\n| `nika catalog` | Provider\u002Fmodel ids + required env vars |\n\n## Procedure\n\n1. Verify readiness: `terminal(command=\"nika --version\")`; install per\n   Prerequisites if missing.\n2. If the task is new, scaffold: `nika new \u003Cfile> --from \u003Ctemplate>`.\n3. Check: `nika check \u003Cfile> --json`. Fix every finding\n   (`nika explain \u003Ccode>`). Do not run an unchecked file.\n4. Preview offline when useful: `nika run \u003Cfile> --model mock\u002Fecho`.\n5. Run with an explicit `--model` and, for any paid model, an explicit\n   `--max-cost-usd`.\n6. For long runs use `background=true` and poll with\n   `process(action=\"poll\"|\"log\")`.\n7. After the run: `nika trace show \u003Ctrace>` + `nika trace verify \u003Ctrace>`\n   (path from the run card); report outputs, actual cost, and the verify\n   verdict to the user.\n\n### Rules\n\n1. NEVER run an unchecked workflow — `nika check` first, every time.\n2. ALWAYS pass `--max-cost-usd` when the model is a paid cloud model.\n3. Prefer local models (`ollama\u002F...`) or `mock\u002Fecho` for drafts; escalate to\n   cloud models only when needed.\n4. Report the final run card honestly: status, actual cost, trace path,\n   `trace verify` verdict.\n5. One workflow file per delegated task; keep files in the user's repo so\n   they are diffable and reusable.\n6. If a run fails, read `nika explain \u003CNIKA-code>` before retrying — do not\n   blind-retry.\n\n## Pitfalls\n\n- `nika run` renders live on a TTY; when piped (Hermes terminal), output can\n  stay quiet until completion — for anything long, prefer `background=true` +\n  poll, then read `nika trace show \u003Ctrace>` for the final card.\n- `nika new` with no `--from` opens a guided TTY flow; in a pipe it fails\n  fast naming the flag — always pass `--from \u003Ctemplate>` when delegating.\n- The budget guard stops NEW admissions: one wide parallel wave can overshoot\n  by that wave's spend. Tighten with `max_parallel:` when the budget is strict.\n- Uncataloged model ids meter as $0 — never rely on `--max-cost-usd` for a\n  custom endpoint model.\n- Workflow `outputs:` are not resolved on a budget stop — per-task values\n  live in the trace (`nika trace outputs`).\n\n## Limitations\n\n- Static checks reduce risk but cannot prove that remote content, shell steps,\n  provider behavior, or generated outputs are safe or correct.\n- Cost caps are not reliable for uncataloged paid models and a parallel wave\n  can overshoot before new work is stopped; require explicit user approval for\n  paid runs and report the actual ledger result.\n- Trace verification proves integrity of the recorded chain, not correctness\n  of the workflow or truth of its outputs.\n\n## Verification\n\nSmoke test (offline, zero keys):\n\n```\nterminal(command=\"nika examples run 01-hello --model mock\u002Fecho\")\n```\n\nSuccess criteria: run completes exit 0 with a final run card · `nika check`\nexits 0 before any real run · `nika trace verify` exits 0 after the run.\n","# Nika Skill\n\nUse [Nika](https:\u002F\u002Fnika.sh) as a deterministic workflow worker orchestrated by\nthe Hermes `terminal` tool. Nika is an open-source (AGPL) Rust engine that captures\na repeatable AI task as a plain-text `*.nika.yaml` file, audits it **before a\nsingle token is spent** (plan, cost floor, secret flows, types), executes it\nagainst local or cloud providers (Ollama\u002Fllama.cpp\u002FvLLM included), and records\na tamper-evident trace.\n\nDivision of labor: **Hermes orchestrates · Nika captures repeatable work as a\ncheckable file and runs it with receipts.** Nika is NOT another coding agent —\nfor autonomous coding, use the `opencode` skill. Delegate to Nika when the\nwork should be *repeatable, budgeted, and auditable*.\n\n## When to Use\n\n- The user asks to run, check, or author a `*.nika.yaml` workflow\n- A task will be repeated (daily digest, triage, ETL, report, multi-step LLM\n  pipeline) — capture it as a workflow instead of re-prompting\n- The user wants a hard cost cap, a cost estimate before running, or\n  receipts\u002Faudit of what ran\n- A pipeline mixes models\u002Fproviders (local + cloud) or mixes LLM steps with\n  shell\u002FHTTP\u002Ffile steps\n- The user wants a run they can replay, verify, or reproduce later\n\n### When NOT to use\n\n- One-off questions or single tool calls — just answer or use a tool\n- Autonomous code implementation\u002Frefactoring\u002FPR review — use the `opencode` skill\n- Interactive back-and-forth tasks — workflows are non-interactive by design\n\n## Prerequisites\n\n- Nika installed: `brew install supernovae-st\u002Ftap\u002Fnika` — other install\n  paths (script, manual download) are documented at https:\u002F\u002Fnika.sh: installing\n  is a human step, not something this skill runs\n- Verify: `terminal(command=\"nika --version\")`\n- Zero keys needed for local\u002Foffline work: `--model mock\u002Fecho` (offline) and\n  `--model ollama\u002F...` (local) run without any API key\n- Cloud providers read standard env vars from the shell;\n  `terminal(command=\"nika doctor\")` diagnoses and prints exact fix commands\n\n## How to Run\n\nProve the toolchain offline first (no key, no network):\n\n```\nterminal(command=\"nika examples run 01-hello --model mock\u002Fecho\")\n```\n\nRun a real workflow — local model first:\n\n```\nterminal(command=\"nika run flow.nika.yaml --model ollama\u002Fqwen3.5:4b\", workdir=\"~\u002Fproject\")\n```\n\nCloud model with a hard budget (always set one for paid models):\n\n```\nterminal(command=\"nika run flow.nika.yaml --model mistral\u002Fmistral-small-latest --max-cost-usd 0.25\", workdir=\"~\u002Fproject\")\n```\n\nPass workflow variables:\n\n```\nterminal(command=\"nika run report.nika.yaml --var city=Paris --var days=7 --max-cost-usd 0.50\", workdir=\"~\u002Fproject\")\n```\n\nLong runs: launch in background and poll — do not block the turn:\n\n```\nterminal(command=\"nika run long.nika.yaml --max-cost-usd 1.00\", workdir=\"~\u002Fproject\", background=true)\nprocess(action=\"poll\", session_id=\"\u003Cid>\")\nprocess(action=\"log\", session_id=\"\u003Cid>\")\n```\n\n### The check-before-run law\n\nNever run a workflow you have not checked. `nika check` is a static pre-flight\n(no tokens spent, no network): plan shape, cost floor, secret-flow analysis,\ntype checks, tool args.\n\n```\nterminal(command=\"nika check flow.nika.yaml --json\", workdir=\"~\u002Fproject\")\n```\n\nFindings carry `NIKA-XXXX` codes that explain themselves via\n`nika explain NIKA-XXXX`. Exit 0 = green, safe to run. Fix findings before\nrunning — never suppress them.\n\n### Authoring a workflow\n\nTurn a repeated task into a file. List templates, then instantiate:\n\n```\nterminal(command=\"nika new --from '?'\")\nterminal(command=\"nika new flow.nika.yaml --from chain\", workdir=\"~\u002Fproject\")\n```\n\n`--from` also accepts plain-words intent. Edit the skeleton (`vars:`,\n`tasks:`, `outputs:`), then **check it**. `nika explain flow.nika.yaml`\nnarrates what it will do, the waves, the cost floor, and what it touches —\nbefore anything runs.\n\nThe artifact you are producing looks like this (checks clean on 0.98):\n\n```yaml\nnika: v1\nworkflow: daily-brief\nmodel: ollama\u002Fqwen3.5:4b\ntasks:\n  - id: fetch\n    invoke:\n      tool: \"nika:fetch\"\n      args: { url: \"https:\u002F\u002Fhn.algolia.com\u002Fapi\u002Fv1\u002Fsearch?tags=front_page\" }\n  - id: brief\n    depends_on: [fetch]\n    infer:\n      max_tokens: 300\n      prompt: |\n        Five bullet points, most signal first: ${{ tasks.fetch.output }}\noutputs:\n  brief: ${{ tasks.brief.output }}\n```\n\nOne file, plain YAML: tasks, an explicit dependency, a bounded model step,\na declared output. That file is what gets checked, run, diffed and reused.\n\n### Cost honesty\n\n- When the workflow prices above the budget, `--max-cost-usd` refuses to\n  start (exit 2, zero tokens) — and since 0.99 the pre-start floor prices\n  the EFFECTIVE model, `--model` override included\n- Mid-run, the ledger stops the workflow the moment real spend crosses the\n  budget: the crossing call completes, nothing new starts, the run fails\n  `NIKA-1704` (exit 1) with spent-vs-budget\n- Estimates use LIST RATES from the vendored public catalog; local · mock ·\n  unpriced work is never blocked\n- A model absent from the catalog meters as $0 — a paid *uncataloged* model\n  runs with no budget protection; prefer cataloged ids (`nika catalog`)\n- Report the cost line from the final run card (the summary block `nika\n  run` prints last — status, cost, trace path) back to the user verbatim\n\n### Receipts and verification\n\nEvery run writes a trace under `.nika\u002Ftraces\u002F` — the run card prints the\ntrace path on its `trace:` line. Both commands take that path (bare\ninvocations are a usage error):\n\n```\nterminal(command=\"nika trace show .nika\u002Ftraces\u002F\u003Crun>.ndjson\", workdir=\"~\u002Fproject\")\nterminal(command=\"nika trace verify .nika\u002Ftraces\u002F\u003Crun>.ndjson\", workdir=\"~\u002Fproject\")\n```\n\n`trace verify` checks the tamper-evidence hash chain: exit 0 intact · 2\nbroken · 3 pre-chain. Also useful: `nika trace outputs` · `nika trace flow` ·\n`nika trace reproduce` · `nika trace export` (OTLP lines).\n\n### Optional: MCP oracle tools\n\nNika also ships a read-only MCP oracle (`nika mcp`) exposing validation and\nlearning tools (`nika_check`, `nika_explain`, `nika_schema`, `nika_examples`,\n`nika_template`, `nika_canon`, `nika_catalog`, `nika_tools`). If the user\nwants those wired into their agent client, point them at the wiring guide —\nhttps:\u002F\u002Fgithub.com\u002Fsupernovae-st\u002Fnika-agents\u002Ftree\u002Fmain\u002Fintegrations\u002Fmcp —\nediting the client's own configuration is the user's step, never this\nskill's. Without the oracle, everything above still works over the terminal;\nrunning workflows stays there regardless, where the budget flags and traces\nlive.\n\n## Quick Reference\n\n| Command | Use |\n|---------|-----|\n| `nika welcome` | What Nika is + what this machine has (offline, exit 0) |\n| `nika new \u003Cfile> --from \u003Ctemplate>` | Scaffold a workflow (`--from '?'` lists) |\n| `nika check \u003Cfile> --json` | Static pre-flight — ALWAYS before run |\n| `nika explain \u003Cfile>` | Narrate: waves, cost floor, touches |\n| `nika run \u003Cfile> --model \u003Cp\u002Fm> --max-cost-usd \u003Cusd>` | Execute with budget |\n| `nika test \u003Cfile>` | Golden test under the mock provider (offline) |\n| `nika trace show\u002Fverify\u002Foutputs\u002Fflow \u003Ctrace>` | Receipts after a run (path from the run card's `trace:` line) |\n| `nika doctor` | Diagnose env\u002Fkeys — prints exact fixes |\n| `nika catalog` | Provider\u002Fmodel ids + required env vars |\n\n## Procedure\n\n1. Verify readiness: `terminal(command=\"nika --version\")`; install per\n   Prerequisites if missing.\n2. If the task is new, scaffold: `nika new \u003Cfile> --from \u003Ctemplate>`.\n3. Check: `nika check \u003Cfile> --json`. Fix every finding\n   (`nika explain \u003Ccode>`). Do not run an unchecked file.\n4. Preview offline when useful: `nika run \u003Cfile> --model mock\u002Fecho`.\n5. Run with an explicit `--model` and, for any paid model, an explicit\n   `--max-cost-usd`.\n6. For long runs use `background=true` and poll with\n   `process(action=\"poll\"|\"log\")`.\n7. After the run: `nika trace show \u003Ctrace>` + `nika trace verify \u003Ctrace>`\n   (path from the run card); report outputs, actual cost, and the verify\n   verdict to the user.\n\n### Rules\n\n1. NEVER run an unchecked workflow — `nika check` first, every time.\n2. ALWAYS pass `--max-cost-usd` when the model is a paid cloud model.\n3. Prefer local models (`ollama\u002F...`) or `mock\u002Fecho` for drafts; escalate to\n   cloud models only when needed.\n4. Report the final run card honestly: status, actual cost, trace path,\n   `trace verify` verdict.\n5. One workflow file per delegated task; keep files in the user's repo so\n   they are diffable and reusable.\n6. If a run fails, read `nika explain \u003CNIKA-code>` before retrying — do not\n   blind-retry.\n\n## Pitfalls\n\n- `nika run` renders live on a TTY; when piped (Hermes terminal), output can\n  stay quiet until completion — for anything long, prefer `background=true` +\n  poll, then read `nika trace show \u003Ctrace>` for the final card.\n- `nika new` with no `--from` opens a guided TTY flow; in a pipe it fails\n  fast naming the flag — always pass `--from \u003Ctemplate>` when delegating.\n- The budget guard stops NEW admissions: one wide parallel wave can overshoot\n  by that wave's spend. Tighten with `max_parallel:` when the budget is strict.\n- Uncataloged model ids meter as $0 — never rely on `--max-cost-usd` for a\n  custom endpoint model.\n- Workflow `outputs:` are not resolved on a budget stop — per-task values\n  live in the trace (`nika trace outputs`).\n\n## Limitations\n\n- Static checks reduce risk but cannot prove that remote content, shell steps,\n  provider behavior, or generated outputs are safe or correct.\n- Cost caps are not reliable for uncataloged paid models and a parallel wave\n  can overshoot before new work is stopped; require explicit user approval for\n  paid runs and report the actual ledger result.\n- Trace verification proves integrity of the recorded chain, not correctness\n  of the workflow or truth of its outputs.\n\n## Verification\n\nSmoke test (offline, zero keys):\n\n```\nterminal(command=\"nika examples run 01-hello --model mock\u002Fecho\")\n```\n\nSuccess criteria: run completes exit 0 with a final run card · `nika check`\nexits 0 before any real run · `nika trace verify` exits 0 after the run.\n",{"name":5,"risk":26,"author":47,"source":48,"license":14,"version":49,"metadata":50,"platforms":67,"date_added":69,"description":7,"source_repo":70,"source_type":31,"prerequisites":71,"license_source":73},"Thibaut Melen (@ThibautMelen) · SuperNovae Studio (github.com\u002Fsupernovae-st)","https:\u002F\u002Fgithub.com\u002Fsupernovae-st\u002Fnika-agents\u002Ftree\u002Fmain\u002Fskills\u002Fnika","1.1.0",{"hermes":51},{"tags":52,"category":59,"homepage":60,"related_skills":61,"requires_toolsets":65},[53,54,55,56,57,58,32],"Workflow","Automation","Deterministic","Cost-Control","Audit","Local-First","autonomous-ai-agents","https:\u002F\u002Fnika.sh",[62,63,64],"opencode","claude-code","codex",[66],"terminal",[41,68],"macos","2026-09-04","supernovae-st\u002Fnika-agents",{"commands":72},[5],"https:\u002F\u002Fgithub.com\u002Fsupernovae-st\u002Fnika-agents\u002Fblob\u002Fmain\u002FLICENSE","f3312ba749561b6159f477dbfc9313c875d40a56",".agents\u002Fskills_archive\u002Fnika","https:\u002F\u002Fgithub.com\u002Fduc-duy-1612\u002FOnline-Delivery-Systems-Web","https:\u002F\u002Fgithub.com\u002Fduc-duy-1612","2026-10-25 23:50:58",{"source":8,"skillName":5,"commands":80,"prompt":90,"zipAvailable":91},[81,84,87],{"tool":82,"command":83},"npx","npx skills add duc-duy-1612\u002FOnline-Delivery-Systems-Web --skill nika",{"tool":85,"command":86},"bunx","bunx skills add duc-duy-1612\u002FOnline-Delivery-Systems-Web --skill nika",{"tool":88,"command":89},"pnpm","pnpm dlx skills add duc-duy-1612\u002FOnline-Delivery-Systems-Web --skill nika","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fduc-duy-1612\u002FOnline-Delivery-Systems-Web\u002Fblob\u002Fmain\u002F.agents\u002Fskills_archive\u002Fnika\u002FSKILL.md",true,[],[94,120,143,165,193,215],{"skillNo":95,"name":96,"nameZh":6,"description":97,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":98,"categories":99,"keywords":103,"publishedAt":42},"01M6Q0FA5FZ410Z33Z3TFQ4DBA","zero-trust","Implement zero-trust network architecture. Configure identity-based access, micro-segmentation, and continuous verification. Use when implementing modern security architectures.","https:\u002F\u002Fgithub.com\u002Fduc-duy-1612\u002FOnline-Delivery-Systems-Web\u002Fblob\u002Fmain\u002F.agents\u002Fskills_archive\u002Fzero-trust\u002FSKILL.md",[100],{"categoryNo":101,"categoryName":102,"taxonomyType":20,"taxonomyTypeDesc":21},"01JBSEEDCAT000000000000011","运维与部署",[104,105,106,107,108,109,96,110,111,112,113,114,31,30,115,26,116,117,118,119],"ZeRO","trust","zero trust","architecture","assessment","verification","segmentation","devops-security","access","configure","only","active","scope","scripts","relevant","implementing",{"skillNo":121,"name":122,"nameZh":6,"description":123,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":124,"categories":125,"keywords":126,"publishedAt":42},"01M6Q0F1YZJW3KX5FBKEVZ8CV4","zcode-delegate","Delegate coding tasks to the Z.AI ZCode CLI only when the user explicitly requests it, while the orchestrator retains review and landing responsibility.","https:\u002F\u002Fgithub.com\u002Fduc-duy-1612\u002FOnline-Delivery-Systems-Web\u002Fblob\u002Fmain\u002F.agents\u002Fskills_archive\u002Fzcode-delegate\u002FSKILL.md",[],[127,128,129,130,131,132,133,134,135,136,137,138,139,114,23,31,140,141,26,142],"delegate","zcode","zcode delegate","review","cli","orchestration","tasks","run","Bash","model","node","read","orchestrator","desktop app","app","agent-orchestration",{"skillNo":144,"name":145,"nameZh":6,"description":146,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":147,"categories":148,"keywords":149,"publishedAt":42},"01M6Q0ETK41VGNMYEXW0T00AJN","youtube-transcript","Fetch YouTube transcripts through DeepAPI or local fallback tooling and save clean text output.","https:\u002F\u002Fgithub.com\u002Fduc-duy-1612\u002FOnline-Delivery-Systems-Web\u002Fblob\u002Fmain\u002F.agents\u002Fskills_archive\u002Fyoutube-transcript\u002FSKILL.md",[],[150,151,152,153,154,64,155,37,156,157,158,159,160,31,30,28,161,162,163,164],"transcript","youtube transcript","youtube","safe","github","research","https","text","fetch","tooling","clean","output","deepapi","youtube transcripts","blob",{"skillNo":166,"name":167,"nameZh":6,"description":168,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":169,"categories":170,"keywords":182,"publishedAt":42},"01M6Q0EPPH2PVMDW40017F8WX0","youtube-transcript-skills","Fetch YouTube video transcripts, search videos\u002Fchannels, browse channels, and extract playlists via the getyoutubetranscript.com API - free tier, no card required.","https:\u002F\u002Fgithub.com\u002Fduc-duy-1612\u002FOnline-Delivery-Systems-Web\u002Fblob\u002Fmain\u002F.agents\u002Fskills_archive\u002Fyoutube-transcript-skills\u002FSKILL.md",[171,174,177],{"categoryNo":172,"categoryName":173,"taxonomyType":20,"taxonomyTypeDesc":21},"01JBSEEDCAT000000000000010","后端与接口",{"categoryNo":175,"categoryName":176,"taxonomyType":20,"taxonomyTypeDesc":21},"01JBSEEDCAT000000000000016","图像与多媒体",{"categoryNo":178,"categoryName":179,"taxonomyType":180,"taxonomyTypeDesc":181},"01JBSEEDCAT000000000000030","开发工程师","PROFESSION","适用职业",[150,151,152,183,153,184,154,64,37,185,186,187,188,158,189,31,30,190,191,192],"youtube transcript skills","gemini","search","api","Video","browse","api-integration","free","card","channels",{"skillNo":194,"name":195,"nameZh":6,"description":196,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":197,"categories":198,"keywords":203,"publishedAt":42},"01M6Q0EAB96PX83W5PTE8A0R8D","youtube-notetaker","Turn YouTube talks into local study notes with slides, transcripts, editable annotations, and a markdown-backed viewer.","https:\u002F\u002Fgithub.com\u002Fduc-duy-1612\u002FOnline-Delivery-Systems-Web\u002Fblob\u002Fmain\u002F.agents\u002Fskills_archive\u002Fyoutube-notetaker\u002FSKILL.md",[199,202],{"categoryNo":200,"categoryName":201,"taxonomyType":20,"taxonomyTypeDesc":21},"01JBSEEDCAT000000000000003","演示文稿",{"categoryNo":175,"categoryName":176,"taxonomyType":20,"taxonomyTypeDesc":21},[152,204,205,131,153,40,206,207,38,37,187,208,209,210,211,30,212,28,213,214],"youtube notetaker","notetaker","official","notes","markdown","slides","annotations","README","study","dair-academy","viewer",{"skillNo":216,"name":217,"nameZh":6,"description":218,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":219,"categories":220,"keywords":222,"publishedAt":42},"01M6Q0E55F0GW5MX48V8HXS8B5","youtube-full","Fetch YouTube transcripts, search videos, browse channels, and extract playlists via TranscriptAPI — no yt-dlp, no Google API key, works from any cloud server.","https:\u002F\u002Fgithub.com\u002Fduc-duy-1612\u002FOnline-Delivery-Systems-Web\u002Fblob\u002Fmain\u002F.agents\u002Fskills_archive\u002Fyoutube-full\u002FSKILL.md",[221],{"categoryNo":172,"categoryName":173,"taxonomyType":20,"taxonomyTypeDesc":21},[223,152,224,225,153,184,185,226,227,186,187,188,158,228,189,31,30,190,229,163],"youtube full","full","google","server","dlp","credits","automatic"]