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 ailiheizi/workflow-to-skill --skill workflow-to-skillgit clone --depth 1 https://github.com/ailiheizi/workflow-to-skillWrote 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/ailiheizi/workflow-to-skill/workflow-to-skill)<a href="https://agentmods.dev/skills/ailiheizi/workflow-to-skill/workflow-to-skill"><img src="https://agentmods.dev/badge/skills/ailiheizi/workflow-to-skill/workflow-to-skill/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/ailiheizi/workflow-to-skill/workflow-to-skill"><img src="https://agentmods.dev/badge/skills/ailiheizi/workflow-to-skill/workflow-to-skill.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 | $0.00105 | $0.03199 |
| Opus 5.5 | $0.00042 | $0.01280 |
| Sonnet 5.5 | $0.00021 | $0.00640 |
| Haiku 4.5 | $0.00011 | $0.00320 |
Grade A, and why
workflow-to-skill 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 — 307 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Workflow To Skill
Create one readable workflow Skill for one coherent reusable outcome. The
generated SKILL.md is an Agent-readable method, not a DAG, DSL, compiled
program, registry entry, or runtime.
The target Harness loads the Skill, runs the Agent, and invokes tools. Existing CLI, MCP, HTTP, API, Dify, n8n, ComfyUI, or other systems perform the real work. Do not implement an executor or copy an external workflow graph into the Skill.
Choose The Authoring Action
This Skill supports two ordinary-language authoring actions. The labels are guidance for the Agent, not commands, metadata, a schema, or a runtime protocol.
- Create: turn a stated goal and confirmed real capabilities into one
candidate workflow
SKILL.md. - Evolve: inspect an existing workflow Skill together with real execution evidence and propose the smallest justified revision. Evolve does not mean automatic self-modification, automatic installation, or automatic promotion.
Use Create when there is no existing Skill to preserve. Use Evolve only when the current Skill revision and evidence from an actual run, test, artifact, or external result can be read. If either is missing, say so instead of inventing a lesson.
Create
goal + real capabilities -> candidate SKILL.md
Execute
Harness + existing tools -> observed result and evidence
Evolve
current Skill + evidence -> root-cause review + smallest candidate diff
External gate
independent replay, tests, or review -> accept, reject, or insufficient evidence
Authoring Method
Adapt this reference flow to the task:
- Clarify the outcome, users, scope, inputs, outputs, side effects, completion criteria, target Harness, and representative success and failure cases.
- Inspect real Harness documentation and available capabilities. Confirm the semantic capability, exact binding, important parameters and results, credential binding names, external workflow identity or version, lifecycle states, and verification signals. Confirm how the target installs, discovers, reloads, and explicitly invokes Skills. Never guess these contracts.
- Present one concise proposal covering the goal, dependencies, inputs, recommended method, decision guidance, posture, recovery, evidence, and test approach. Label user-confirmed choices, proposed safe defaults, and unresolved material decisions.
- Ask for review only where a choice could materially change the goal, target, scope, cost, authority, side effects, or success criteria. Harmless formatting and low-risk defaults may remain Agent judgment.
- Write or update the candidate
SKILL.mdafter material decisions are confirmed or covered by an explicit unattended policy. Otherwise return an incomplete proposal and state what prevents finalization. - Re-read and structurally validate the candidate. When authorized, install it through the Harness's existing mechanism, confirm the intended candidate was discovered rather than a stale or conflicting copy, and explicitly invoke it through the ordinary Skill path.
- Exercise a safe representative success case and material stop or failure cases that can be tested without disproportionate risk. Separate observed behavior from untested dependencies and lifecycle claims.
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 Changed · +101 lines · +1 tokens per session 5a658e261542
- 2mo ago First seen · 206 lines · 104 tokens per session scan A e991fe3c3e91
workflow-to-skill is a skill published in the GitHub repository ailiheizi/workflow-to-skill (2 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 105 tokens to every session and 3,199 once invoked, about $0.0004 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
agent-loop
Detect requests for iterative autonomous agent loops and route to the appropriate loop executor.
cross-task-learner
Enable agent loops to learn from similar past tasks and share patterns across loops.
aiwg-language-map
AIWG addons + extensions language map — categories, curated discover phrases, and per-bundle pointers covering everything beyond the framework quickrefs.
sdlc-accelerate
End-to-end SDLC ramp-up from idea to construction-ready with automated phase transitions.
steward
Provide provider capability awareness and route commands by answering what the current provider supports natively vs must emulate.
infer-completion-criteria
Infer measurable completion criteria for an agent-loop task from project docs, code, and AIWG standards when the user has not supplied --completion explicitly.