Repo Sweeper — Development agent for Claude Code
Read-only mechanical sweeps — PR and CI status across many PRs, worktree/branch leftovers, issue states, path and link existence, grep hit counts.
How to install Repo Sweeper
Installs to ~/.claude/agents/mrveiss-autobot-ai-repo-sweeper.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/mrveiss/AutoBot-AI/HEAD/.claude/agents/repo-sweeper.md -o ~/.claude/agents/mrveiss-autobot-ai-repo-sweeper.md Restart Claude Code, or start a new session, for it to be picked up.
What Repo Sweeper does
name: repo-sweeper description: Read-only mechanical sweeps — PR and CI status across many PRs, worktree/branch leftovers, issue states, path and link existence, grep hit counts. Use when the same deterministic check repeats across many targets, or when raw output would flood the main context. Returns observations in a fixed format and never draws conclusions. Do NOT use for reviews, judgement calls, or anything that writes. model: haiku tools: Read, Grep, Glob, Bash
You run **mechanica
Alternatives in Development
- Dashclaw Drift Auditor — Audits DashClaw's drift-prone hardcoded counts and version stamps — SDK method counts (Node/Python), MCP tool/ 297 ★
- Clawmem Curator — Maintains ClawMem memory health — lifecycle triage (pin/snooze/forget), retrieval health checks, dedup sweeps 201 ★
- Loom Daemon — Loom Daemon operator - Tier 2 daemon-mode surface that observes the running Rust loom-daemon binary via MCP to 51 ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Indicator Sweeper
Sweeps open indicators of an OSINT case against the live web and returns graded, sourced findings (or an expli
Haiku Worker
Executes one already-decided step from a ticket's plan. Dispatched by execute when a step is mechanical and re
Fk Benchmarker
Owns the measurement protocol - noise floor, warm-up, clock ramp, repeats, CUDA sync, peak VRAM, launch counts
Doc Link Checker
Validates documentation links — internal file/anchor existence plus external URL checks. Returns a broken/redi
Orch Explorer
Read-only codebase search. Use for sweeps across many files — where something lives, how it is used, what touc
Agent Performance Updater
Tier-1 Trade Evaluation. Computes per-agent (hit_rate, sharpe, pnl, n_samples) rows for the agent_performance
Related Skills
Kb Index
Hybrid (vector + BM25) search over many markdown repos in one SQLite file, served over MCP for coding agents:
Rename Code
Safely rename a symbol/string across the codebase — grep all refs, scope-check, apply, verify with build + zer
Dsh Fast
Read-only performance diagnostics for DeepSeek Harness: session load/restore timing, spill-hit counts, compact