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 agentmods add commands/marioepkone/pipelineiq/audit-implementationgit clone --depth 1 https://github.com/MarioEpkOne/PipelineIQWrote 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/commands/marioepkone/pipelineiq/audit-implementation)<a href="https://agentmods.dev/commands/marioepkone/pipelineiq/audit-implementation"><img src="https://agentmods.dev/badge/commands/marioepkone/pipelineiq/audit-implementation.svg" alt="Measured on agentmods" 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 | $0.00000 | $0.02310 |
| Opus 5 | $0.00000 | $0.01155 |
| Sonnet 5 | $0.00000 | $0.00462 |
| Haiku 4.5 | $0.00000 | $0.00231 |
Grade A, and why
audit-implementation 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 5d 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 — 198 lines — stays where its author put it; the contents beside it link to each section on GitHub.
POST-IMPLEMENTATION AUDIT — analyzes what the implementing agent did wrong and why, proposes specific changes to CLAUDE.md / impl.md / impl-plan.md for approval, and appends new patterns to learnings.md.
The user's request is: $ARGUMENTS
Purpose
This skill is a process autopsy, not a QA check. The goal is to understand why failures happened so the skills and rules can be improved for future implementations. Run this after every /impl. For auditing whether an implementation plan conforms to the impl-plan skill spec, use /audit-plan instead.
Re-audit mode (after /fix): If this audit is being run after a /fix session, limit scope to: (1) verify each change the fixer stated it made in its working log, and (2) confirm no regressions were introduced. A full independent evaluation and complete failure taxonomy are not needed — skip Phase 2 (the blind pass) entirely and read all inputs normally, then condense Phase 3 to only the fixer's stated changes. Note "Re-audit — scoped to fixer's stated changes" at the top of the audit document.
Phase 1 — Find Inputs
Find the working log:
- If an argument was given: search
Working Logs/for a file whose name contains the argument (case-insensitive). If multiple match, list them and ask. - If no argument: use the most recently modified
.mdfile inWorking Logs/. - If none found: tell the user and exit.
Read only the working log's header fields (e.g. **Impl plan:**) — enough to follow the chain to the spec. Do not read the working log body ("Changes Made", "Deviations", "Errors", verification narrative) yet; that is deferred to Phase 3 so it cannot anchor your independent verdict.
Re-audit exception: In Re-audit mode (after
/fix), there is no blind pass — read the working log in full now, normally.
Follow the chain:
- From the working log's
**Impl plan:**header field: find the impl plan inImplementation Plans/and read only its**Spec:**header field (not the full plan — the plan body is deferred to Phase 3 along with the working-log body). - From the impl plan's
**Spec:**header field: find and read the spec in full fromspecs/.
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.
- 5d ago First seen · 198 lines · 0 tokens per session scan A fd1fbb7ad713
audit-implementation is a command published in the GitHub repository MarioEpkOne/PipelineIQ (2 stars, last pushed 2mo ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 2,310 tokens. 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 commands, from other repositories
promote
Refine scv/raw/ material into a scv/promote/ - - / folder with PLAN.md + TESTS.md. Optionally updates the docs knowledge graph. Interactive; no files written without user approval. Use whenever the user wants a change planned, specced, or broken down before implementation — 'plan this', 'write a spec', 'what should we…
work
Implement a scv/promote/ / plan. Reads PLAN.md + TESTS.md, proposes/loads Related Documents as needed, runs the tests, and optionally archives on success. Use whenever the user wants an existing plan implemented — 'build it', 'implement the plan', 'let's do it' — not only when they type /scv:work. Never implement a…
help
Show SCV workflow + diagnose project + recommend next step. With an argument, talk through an idea OR search the archive for past work — SCV picks the right mode from your wording. Use whenever the user is thinking out loud about a change, asks what to do next, asks what SCV can do, or wants to find past work — not…
codegen
TDD-first codegen for a scv/promote/ / plan. Verifies TESTS fail first (Red), then iteratively writes the minimum code to pass each case (Green). Hands off archive/PR/regression to /scv:work. Use whenever the user wants a plan implemented test-first — 'TDD', 'tests first', 'red-green' — not only when they type…
deck
Turn a markdown planning doc into a spec-grade 기획서 HTML — a buildless, self-contained document by default, or an on-screen slide presentation when you ask for one. Deterministic md→deck transform + a quality/gap lint; the raw markdown travels with it. Use whenever the user wants a planning document turned into…
regression
Run accumulated regression across scv/archive//TESTS.md (and optionally promote/) with supersede/obsolete skip graph. On failure, triage each slug via AskUserQuestion (regression / obsolete / flaky). Use whenever the user asks whether past features still pass, or before archiving anything — 'run the tests', 'is…