Retroactive — Development skill for Claude Code
Improve the Harness after a bug or miss, so the same class of problem can't recur.
How to install Retroactive
Installs to ~/.claude/commands/cixin-dev-ai-layer-template-retroactive.md
mkdir -p ~/.claude/commands && curl -fsSL https://raw.githubusercontent.com/cixin-dev/ai-layer-template/HEAD/.claude/commands/retroactive.md -o ~/.claude/commands/cixin-dev-ai-layer-template-retroactive.md Restart Claude Code, or start a new session, for it to be picked up.
What Retroactive does
description: Improve the Harness after a bug or miss, so the same class of problem can't recur argument-hint:
Retroactive
**Input**: $ARGUMENTS
This is the **System Evolution** outer loop (see [`../skills/system-evolution/SKILL.md`](../skills/system-evolution/SKILL.md)). Do not stop at patching the surface code — fix the Harness so the problem can't recur.
You are reached through one of two deterministic entry points (ADR-0010), never from
Alternatives in Development
- Contributing — Thanks for your interest in contributing to Marketing Skills 13.8k ★
- Oob — Out-of-band orchestrator — confirm BLIND SSRF/XXE/SQLi/RCE/Log4Shell by correlating interactsh callbacks to th 4.5k ★
- AWorld — Search, understand, reproduce, and improve an idea with ease 1.2k ★
Full documentation available on GitHub
View Source RepositoryRelated Skills
Atharvan Dtd
remedy-first for a bug class; each remedy is a charm (the fix) and a rite (the verification), contraindication
Docs Refresh
Review and improve docs against the live harness surface
Craft Reflect
Apply captured learnings to improve the Craft harness and project patterns. Converts .learnings.yaml into perm
Pair Verify
Use after you fix a bug and before you claim it fixed — it spawns ONE skeptic sibling session via CCC that mus
Report Draft
Draft a HackerOne bug bounty report from a validated finding. Only invoke this after a full end-to-end PoC exi
Deep Solve
Solve a hard self-contained problem via the deep-solve convergence harness
Related Agents
Agent Tuner
Reflexion-style agent tuner. Given evidence of a recurring failure by a specific agent, proposes the smallest
Harness Optimizer
Analyze and improve the local agent harness configuration for reliability, cost, and throughput.
Swe Dev Structured Problem Solver
A research-backed problem-solving harness for software engineering. Uses Wankat and Oreovicz's engineering loo