Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add shenchangmin/cairnkit --skill workflow-orchestratorgit clone --depth 1 https://github.com/shenchangmin/cairnkitWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/shenchangmin/cairnkit/workflow-orchestrator)<a href="https://agentmods.dev/skills/shenchangmin/cairnkit/workflow-orchestrator"><img src="https://agentmods.dev/badge/skills/shenchangmin/cairnkit/workflow-orchestrator/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/shenchangmin/cairnkit/workflow-orchestrator"><img src="https://agentmods.dev/badge/skills/shenchangmin/cairnkit/workflow-orchestrator.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00018 | $0.02300 |
| Opus 5.5 | $0.00007 | $0.00920 |
| Sonnet 5.5 | $0.00004 | $0.00460 |
| Haiku 4.5 | $0.00002 | $0.00230 |
Grade A, and why
workflow-orchestrator scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 182 lines — stays where its author put it; the contents beside it link to each section on GitHub.
workflow-orchestrator
You are a thin driver. Do not retain workflow state in your head or decide transitions:
the cairn binary is the only writer of .cairnkit/STATE.yaml. On every iteration, read the
state, do exactly the current stage's work, then ask cairn to advance.
The CLI you drive (from the host project root)
cairn --root . state show # current state (JSON)
cairn --root . state resume # {stage, paused}
cairn --root . state advance # next stage, gate-checked
cairn --root . state set-path-mode <full|lite|single|tooling>
cairn --root . state approve-clarify # clear a CLARIFY pause
cairn --root . state fail --stage <BUILD_VERIFY|E2E_VERIFY>
cairn --root . state unblock # after a human resolves a blocked run
cairn --root . kb build-index
cairn --root . kb query --stage <S> --budget 300 [--domain <d>]
cairn --root . kbrepo sync-in | sync-out
Exit codes: 0 success; 2 usage/configuration error; 3 gate refused; 4 corrupt state.
Never edit .cairnkit/STATE.yaml by hand.
Stage map
| stage | role | expected artifact |
|---|---|---|
| ANALYSE_PRODUCT | product |
01-product.md |
| ANALYSE_TECH | tech |
02-tech.md |
| ARCHITECT_BACKEND | architect-be |
03-arch.md |
| ARCHITECT_FRONTEND | architect-fe |
04-arch-fe.md |
| IMPLEMENT | dev |
05-implement.md |
| BUILD_VERIFY | verify |
06-build.md |
| VISUAL_REVIEW | visual |
07-visual.md |
| E2E_VERIFY | verify |
08-e2e.md |
| TEST | verify |
09-test.md |
| ARCHIVE | archiver |
10-archive.md |
| INIT, INTENT_GATE, CLARIFY_*, DONE | — | — |
The expected path for a role stage is
docs/workflows/<run-id>/<expected artifact>.
Route at INTENT_GATE
Classify the original user request yourself. Do not interpolate request text into a shell command; the model classification is the routing decision.
single— a single-point/configuration/governance/docs change with no analysis or design.lite— a backend or non-UI product change. It skips frontend architecture and visual review, but still includes E2E verification.tooling— a non-UI tooling or governance deliverable. It follows the lite analysis/design path but also skips E2E verification; this is an engine-enforced route, never a stub E2E artifact.full— a feature with frontend/UI work, including frontend architecture and visual review.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 1mo ago First seen · 182 lines · 18 tokens per session scan A 07d6a2e01c3e
workflow-orchestrator is a skill published in the GitHub repository shenchangmin/cairnkit (1 stars, last pushed 2mo ago), licensed MIT. It adds 18 tokens to every session and 2,300 once invoked, about $0.0001 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
tasks
View and manage the task stack and processing queue. Shows pending work, active tasks, completed items, and queue state. Triggers on "/tasks", "show tasks", "what's pending", "task list", "queue status".
llm-wiki
Build and maintain an LLM-curated knowledge base from papers, articles, transcripts, notes and project findings. Ingest sources into linked Markdown, answer questions with citations, lint or upgrade a wiki, and preserve useful synthesis across sessions and agents. Also capture verified task experience, consolidate…
reduce
Extract structured knowledge from source material. Comprehensive extraction is the default — every insight that serves the domain gets extracted. For domain-relevant sources, skip rate must be below 10%. Zero extraction from a domain-relevant source is a BUG. Triggers on "/reduce", "/reduce [file]", "extract…
rethink
Challenge system assumptions against accumulated evidence. Triages observations and tensions, detects patterns, generates proposals. The scientific method applied to knowledge systems. Triggers on "/rethink", "review observations", "challenge assumptions", "what have I learned".
graph
Interactive knowledge graph analysis. Routes natural language questions to graph scripts, interprets results in domain vocabulary, and suggests concrete actions. Triggers on "/graph", "/graph health", "/graph triangles", "find synthesis opportunities", "graph analysis".
next
Surface the most valuable next action by combining task stack, queue state, inbox pressure, health, and goals. Recommends one specific action with rationale. Triggers on "/next", "what should I do", "what's next".