Pipeline Clerk — Development agent for Claude Code
Caretake the delivery pipeline: sweep its open pull requests for dead fires and repair them, straighten a stuck item's labels, run the automated code-review round, and take one new finding into a skel.
How to install Pipeline Clerk
Installs to ~/.claude/agents/akurganow-ai-plugins-pipeline-clerk.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/Akurganow/ai-plugins/HEAD/.claude/agents/pipeline-clerk.md -o ~/.claude/agents/akurganow-ai-plugins-pipeline-clerk.md Restart Claude Code, or start a new session, for it to be picked up.
What Pipeline Clerk does
name: pipeline-clerk description: "Caretake the delivery pipeline: sweep its open pull requests for dead fires and repair them, straighten a stuck item's labels, run the automated code-review round, and take one new finding into a skeleton branch and draft pull request. Use for the pipeline's scheduled run." model: inherit skills: [pipeline-law, pipeline-clerk]
You are the Pipeline Clerk of this repository's delivery pipeline.
Your role is the `pipeline-clerk` skill, preloaded above. I
Alternatives in Development
- Review Concurrency — Reviews code changes for concurrency defects found by multi-thread interleaving reasoning: race conditions, vi 432 ★
- Human Escalation Agent (Stuck Handler) — You are the STUCK AGENT - the MANDATORY human escalation point for the entire system 316 ★
- Ontology Term Lookup — Use this agent when you need to find ontology terms by their textual labels or descriptions using the OLS4 MCP 163 ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Tracker Clerk
Close the issues of this repository whose findings are provably gone, whose claim the court called a duplicate
Maintainability Reviewer
Read-only maintainability reviewer for the forgeward gate. Fires on every diff that touches code. Audits dead
Label Classifier
Assigns type:, priority:, effort: labels to a backlog item, each with one-line reasoning. Returns 'unclear' pe
Engine Grounding Scout
Use when you are about to run the memory-recall sweep or a knowledge-graph impact traversal and the sweep will
Proj Work
Use this agent when creating or executing work items during automated development cycles. Typical triggers inc
ChatGPT PR Review
Coordinates ChatGPT PR review round-by-round in a dedicated session (manual / automated / parallel modes). Rea
Related Skills
Audit End Sprint
End-of-sprint code audit — run 9-dimension sweep (security, quality, tests, dead code, TODO/FIXME, perf, docs,
Devil Mode
Parallel adversarial sweep — fires prompt-injection-hunter + supply-chain-auditor + secret-hunter + backdoor-h
Unstick
Diagnose and fix a stuck card — clears stale state, labels, worktrees