A coordinator for carrying out a documented workflow contract in its own Git worktree, an isolated working copy of a repository. It delegates implementation and review, verifies each stage, opens a guarded draft pull request, and runs user acceptance testing.

Use it when a task already has a workflow contract and must pass its gates. It coordinates implementation, review, fixes from comments, draft-PR creation, and final checks.

It makes multi-stage development auditable and prevents a task from being declared finished before required verification and review evidence exists.

wf-impl — Skill for Claude Code by tomsej/pi-ext

Install

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.

Any agent

npx skills add tomsej/pi-ext --skill wf-impl

Clone the repo

git clone --depth 1 https://github.com/tomsej/pi-ext

Made for: Claude Code.

Wrote 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.

README.md

[![agentmods](https://agentmods.dev/badge/skills/tomsej/pi-ext/wf-impl/github.svg)](https://agentmods.dev/skills/tomsej/pi-ext/wf-impl)

Your own site

<a href="https://agentmods.dev/skills/tomsej/pi-ext/wf-impl"><img src="https://agentmods.dev/badge/skills/tomsej/pi-ext/wf-impl/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.

Your own site · 80×15

<a href="https://agentmods.dev/skills/tomsej/pi-ext/wf-impl"><img src="https://agentmods.dev/badge/skills/tomsej/pi-ext/wf-impl.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 89 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 2,668 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe. Third-party audits

NVIDIA SkillSpector warn 7 Sept 2026

SkillSpector: 3 findings, up to medium

These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →

medium Agent Snooping · line 76
Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.

Fix: Remove all code or instructions that list or read other skills' files or directories. Skills should operate independently; cross-skill access is a privilege escalation.
medium Agent Snooping · line 104
Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.

Fix: Remove all code or instructions that list or read other skills' files or directories. Skills should operate independently; cross-skill access is a privilege escalation.
medium Agent Snooping · line 105
Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.

Fix: Remove all code or instructions that list or read other skills' files or directories. Skills should operate independently; cross-skill access is a privilege escalation.

How audits are shown

Origin original No closer match found in the catalogue.

Token cost

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	$0.00089	$0.02668
Opus 5.5	$0.00036	$0.01067
Sonnet 5.5	$0.00018	$0.00534
Haiku 4.5	$0.00009	$0.00267

Measured 1mo ago against content hash 106e57ed2755, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

wf-impl 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.

skills/wf-impl/SKILL.md · 153 lines

How it starts

The opening of the file, as written. The whole thing — 153 lines — stays where its author put it; the contents beside it link to each section on GitHub.

/wf-impl — dirigent kontraktu (worktree session)

Diriguješ, produkční kód nepíšeš. Každá fáze = delegace; přechody řídí VÝHRADNĚ exit kódy wf-gate — „hotovo" od agenta neznamená nic. Nikdy nespouštěj sc worktree create. Foreground příkazy v gtimeout <s>, dlouhé běhy přes bg_start — nikdy neomezený foreground watch.

GATE = node ~/Workspace/pi-ext/wf/wf-gate.mjs SPEC = absolutní cesta ke contract.md z tvého tasku (kontrakt má vlastní adresář ~/Workspace/specs/<projekt>/<název>/ i pro další artefakty). Když soubor chybí, zastav a reportuj — kontrakt nevymýšlej ani nehledej v repu.

Extension wf-gate blokuje gh pr create, dokud v .wf/receipts.jsonl není průchozí FULL verify a review attest na aktuálním čistém HEADu; gh pr merge je blokovaný vždy. Guard neobcházej, splň ho — reason říká, který receipt chybí.

Fáze 0 — orientace (a zároveň resume)

Přečti SPEC; spusť GATE begin SPEC (nabije guard, spotřebuje launch claim).
GATE agents SPEC --json → impl a review — modely nikdy neodhaduj z YAML.
GATE verify SPEC preflight --json — levná kontrola prostředí; fail zastaví práci před buildem. Starší kontrakt bez preflightu jen viditelně přeskoč.
Hotové fáze odvoď z reality (git log na branchi, receipts, gh pr view --json state); pokračuj od první nedokončené.

Fáze 1 — implementace

Jeden subagent_spawn s harness, model a reasoning_effort z impl, working_dir = tenhle worktree. Dirigent drží krátký orchestration kontext a do produkčního kódu nesahá. Zadání:

Implementuj kontrakt na <SPEC> striktně TDD po skupinách kritérií:

Z kontraktu čti sekce Akceptační kritéria, Strategie testování, Scope, Non-goals a Přístup; frontmatter, verify a review plán jsou věc dirigenta.

Rozděl kritéria na soudržné skupiny (jeden modul / jedno chování, typicky 3–5 kritérií, ideálně do osmi skupin) a rozpad napiš do reportu.

Na každou skupinu jeden cyklus: RED (jeden test soubor, testy celé skupiny padají ze správného důvodu) → GREEN (minimální kód) → refactor jen na zeleném. Cíl 1–2 test případy na kritérium (tabulka případů = jeden test); víc jen kde kritérium jinak nepřibiješ.

Kritérium platné už na baseline (zachované chování, regrese) RED nemá: napiš charakterizační test, ukaž ho zelený před i po změně, v reportu označ jako regresní. Nikdy kvůli RED nerozbíjej produkční kód.

V RED/GREEN pouštěj jen dotčený test soubor, jedním příkazem krátkým na zeleno i vypovídajícím na červeno (node --test <soubor> 2>&1 | tail -30 nebo ekvivalent). Stejný test nikdy dvakrát kvůli jinému reporteru — jeden dražší výstup stačí. Celou sadu nepouštěj — tu vlastní gate.

Kontext drž krátký: nevypisuj právě zapsané soubory, nečti znovu, co už máš.

Drž se sekce Scope, respektuj non-goals a vyloučené přístupy. Test nikdy neoslabuj, aby prošel.

Commit po každé skupině (Conventional Commits); branch nepřejmenovávej.

Žádné review, žádní revieweři. Nikdy gh pr create ani gh pr merge — endgame vlastní dirigent (u codex harnessu je tahle věta jediná obrana, guard tam není).

Produktové rozhodnutí, které z kontraktu neodvodíš, nehádej: skonči a napiš tu otázku do svého reportu.

Read the full file on GitHub · 153 lines

Changes

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 · 153 lines · 89 tokens per session scan A 106e57ed2755

Subscribe to this mod's changes

wf-impl is a skill published in the GitHub repository tomsej/pi-ext (79 stars, last pushed 1mo ago), licensed MIT. It adds 89 tokens to every session and 2,668 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 2026-08-30.

Git and workflow Planning

wf-impl

Getting it into your agent

Wrote this? Show the measurements

What it costs to keep this loaded

Grade A, and why

Nothing flagged

How it starts

/wf-impl — dirigent kontraktu (worktree session)

Fáze 0 — orientace (a zároveň resume)

Fáze 1 — implementace

What this file has done since we first saw it

Other skills, from other repositories

merge-batch

prowler-commit

gh-auth-isolation

comet-github

github-skill

release-helper