Base Comparer — Development agent for Claude Code
The regression gate, in a clean context: judges every changed bullet of a variant blind against the base line it replaces, for this JD, and reverts any line that does not win; the base ships unless th.
How to install Base Comparer
Installs to ~/.claude/agents/kahl-d-jaja-base-comparer.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/Kahl-d/jaja/HEAD/.claude/agents/base-comparer.md -o ~/.claude/agents/kahl-d-jaja-base-comparer.md Restart Claude Code, or start a new session, for it to be picked up.
What Base Comparer does
GENERATED by scripts/kernel/sync.py from agents/base-comparer.md - do not edit
name: base-comparer description: "The regression gate, in a clean context: judges every changed bullet of a variant blind against the base line it replaces, for this JD, and reverts any line that does not win; the base ships unless the variant is a clear win." tools: Read, Grep, Glob, Bash, Write model: opus skills: [base-compare] disallowedTools: Edit, Agent effort: medium maxTurns: 30 denyCommands:
- gmail_s
Alternatives in Development
- Performance Guardian — Automated performance testing, regression detection, and optimization recommendations 108 ★
- Fixer Sonnet High — Fix-only agent, tier 2. Fixes known-cause findings that span modules or a seam, or known probe classes, with r 77 ★
- Bench — API performance benchmarking — latency profiling, throughput testing, performance regression detection 73 ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Prospect Triage
Scores every new or changed prospect for warmth, fit, sponsorship (a weighed signal; citizenship and clearance
Council Opus
Council seat - intent and edge cases. Judges what the owner meant but did not write, evidencing every ask by r
Resume Reverify
Audits the fix pass of the tailoring pipeline. Verifies every field the fixer changed still traces to the bank
Stamity Fixer
Applies minimal fixes for the Critical and Warning findings of one review round, answers findings it judges wr
Roadmap Pilot Agent
Utiliser pour comparer en continu l'avancement effectif de la roadmap à la trajectoire prévue. Produit une not
Review Outcomes
Periodic post-ship outcome review — grade shipped merges against the fixes and reverts that followed, and agai
Related Skills
Last Pass
Stage 4 last pass - fresh-eyes regression against ALL prior milestones (_firstinitial through _finalinitial),
Settled Computer
Fast computer use for AI agents: event-driven settling MCP server - replaces time.sleep between clicks, waits
Writing Craft
A blind tested skill for better creative writing. Nine signals three AI judges named after scoring 50 samples