Failure Log — Development skill for Claude Code
Append a FAIL-XXX entry to .memory/failures/log.md when a hypothesis fails verification.
How to install Failure Log
Installs to ~/.claude/commands/extraforthesystem-beep-knowledge-harness-failure-log.md
mkdir -p ~/.claude/commands && curl -fsSL https://raw.githubusercontent.com/extraforthesystem-beep/knowledge-harness/HEAD/.claude/commands/failure-log.md -o ~/.claude/commands/extraforthesystem-beep-knowledge-harness-failure-log.md Restart Claude Code, or start a new session, for it to be picked up.
What Failure Log does
description: Append a FAIL-XXX entry to .memory/failures/log.md when a hypothesis fails verification.
When a fix or hypothesis fails SRE verification (after rollback attempt 3), **or** when the shell hook reports the same command failed 3×:
- Read
.memory/failures/log.md - Append a new entry:
## FAIL-XXX: Title
- **Date:** YYYY-MM-DD
- **Hypothesis:** ...
- **Result:** Failed — ...
- **Lesson:** ...
- Run
python scripts/harness-sync.py
Confirm FAIL id assigned
Alternatives in Development
- Ideation — Ideation — Literature Search & Hypothesis Generation 613 ★
- Verify Theme — Run the full Graphite quality gate (drift, tokens, contrast, YAML) and report pass/fail 452 ★
- MCP Edit — tilth_write replaces the host Edit and Write tools 349 ★
Full documentation available on GitHub
View Source RepositoryRelated Skills
Close Phase
Close a project phase — tick the checklist, update plan.md status, archive acceptance results, append a SESSIO
ParcelRound
Splitting work across several coding agents is easy; the seam between them is what fails. A method for the sea
Fork Pov
Fork pov.md for installer taste, or append a one-liner to gotchas.md after a real agent failure.
App Learn
Mine a shipped app's conventions into the living House Knowledge Base — adds net-new learnings to knowledge/.m
Update Status
Append a new stage to an existing job application's status in the tracking log
Allow Asks
Review the permission-prompt log, propose allow-rule patterns, and append approved ones to ~/.claude/settings.
Related Agents
Flaky Test Isolator
USE WHEN a test intermittently fails on unchanged code. Runs it N times sequentially, captures pass/fail + std
Eval Failure Analyzer
Analyze Logic-Lens benchmark/eval failures. Use after running content-evals, or when pointed at a skills-works
Cr Silent Failure Hunter
Hunts silent failures in the PR diff — empty or overly broad catch blocks, fallbacks that mask errors, excepti