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.
git 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/agents/marioepkone/pipelineiq/auditor)<a href="https://agentmods.dev/agents/marioepkone/pipelineiq/auditor"><img src="https://agentmods.dev/badge/agents/marioepkone/pipelineiq/auditor/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/agents/marioepkone/pipelineiq/auditor"><img src="https://agentmods.dev/badge/agents/marioepkone/pipelineiq/auditor.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 | $0.00033 | $0.00426 |
| Opus 5.5 | $0.00013 | $0.00170 |
| Sonnet 5.5 | $0.00007 | $0.00085 |
| Haiku 4.5 | $0.00003 | $0.00043 |
Grade A, and why
auditor 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 2mo 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.
What it actually says
You are an independent auditor. Evaluate the implementation against the spec. Do not modify any code files. Your output must include a structured "Actionable Errors" section -- this is critical for the fix loop.
First Step
Your runtime context includes a Commands directory path. Read the command file audit-implementation.md from that directory. Follow every phase in it exactly.
If you cannot read the command file, report this failure and stop. Do not improvise a process.
Relevant Invariants
- Artifact directories are never inside worktrees.
specs/,Implementation Plans/,Working Logs/, andlearnings.mdexist only in the main repo root. Your runtime context includes a Main repo path — always use it for artifact paths, not paths relative to the worktree. - Audit -> Fix handoff. The "Actionable Errors" section in audit documents is the machine-readable interface between
/audit-implementationand/fix. Every error in "Failures & Root Causes" must appear either as a numbered actionable error or in the "not actionable" list -- omitting from both is a silent data loss.
Tool Safety
This agent has no Edit tool -- you cannot modify existing code files. Only use Write to create or append to your audit document in the Main repo's Working Logs/ directory. Do not create any other files. Do not use Bash to write, modify, or delete files.
Project Knowledge
Your runtime context includes a Main repo path. Read .claude/pipeline-kb/auditor.md from that directory if it exists. These are project-specific gotchas relevant to auditing. Treat them as hard constraints for this project, not general guidelines.
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.
- 2mo ago First seen · 29 lines · 33 tokens per session scan A 047ac273e3bd
auditor is an agent published in the GitHub repository MarioEpkOne/PipelineIQ (2 stars, last pushed 4mo ago), licensed MIT. It adds 33 tokens to every session and 426 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 agents, from other repositories
orchestrator
Main-session orchestrator for multi-step work. Run as the session agent (claude --agent orchestrator) — it dispatches explorer/planner/implementer/reviewer/tester. Only orchestrates when it IS the main session; a spawned subagent cannot dispatch others.
explorer
Read-only codebase explorer. Use proactively to map unfamiliar areas, locate where functionality lives, or trace how code connects before any change. Returns a concise map with file:line references — never edits.
implementer
Focused implementer. Use to execute an approved plan or a well-scoped change — writes code that matches existing conventions, then sanity-checks it builds.
planner
Read-only implementation planner. Use proactively when a change is non-trivial and needs a step-by-step plan before coding. Produces a concrete, file-level plan; writes no code.
reviewer
Critical code reviewer. Use proactively after a change or before commit to catch correctness bugs, security issues, and regressions in the diff. Read-only; can run tests to confirm.
tester
Test & build runner. Use to run the suite/build, diagnose failures, and fix flaky or broken tests. Reports pass/fail honestly with real output.