Kjell Tore Guttormsen
7b3b487d26
The recall+precision halves of the CA-OPT hybrid motor for the three
mechanism-fit cases the deterministic OPT scanner (2a) deliberately skips:
lifecycle→hook (BP-MECH-001), unscoped path-specific→rule (BP-MECH-002),
absolute "never"→permission (BP-MECH-004).
New:
- scanners/lib/lens-prefilter.mjs — cheap, recall-oriented line scan of the
CLAUDE.md body; detector names mirror the register lensCheck fields; skips
fenced code, gates the path class on an instruction verb. Pure + 13 tests.
- scanners/optimize-lens-cli.mjs — discovery + OPT scanner + pre-filter; attaches
only the CONFIRMED register entry to each candidate (unverifiable → dropped,
Verifiseringsplikt); emits {deterministic, candidates, register, counts}.
- agents/optimization-lens-agent.md — opus precision gate (7th agent, orange):
reads the real CLAUDE.md, drops low-confidence candidates, keeps only genuine
opportunities, cites register id + source.
- commands/optimize.md — /config-audit optimize orchestrates pre-filter→agent→report.
Agent-driven → deliberately NOT byte-stable (own command, outside the snapshot
suite). No new orchestrated scanner → scanner count stays 15. Counts: agents
6→7, commands 18→19, suite 1055→1068. Self-audit A/A unchanged, readmeCheck
passed (clean HOME). Plan: docs/v5.7-optimization-lens-plan.md.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
4.7 KiB
4.7 KiB
| name | description | allowed-tools | model |
|---|---|---|---|
| config-audit:help | Show all available config-audit commands | Read, Bash | sonnet |
Config-Audit: Help
Getting Started
Just run /config-audit — it auto-detects your project scope and runs a full audit. No setup needed.
The default output is written in plain language: each finding is grouped by impact ("Configuration mistake," "Conflict," "Wasted tokens," "Missed opportunity," "Dead config") and led with an urgency phrase ("Fix this now," "Fix soon," "Fix when convenient," "Optional cleanup," "FYI").
If you prefer the v5.0.0 verbatim output (technical IDs, raw severity, no plain-language wording), pass --raw to any command — it's threaded through every CLI in the toolchain. Use the Read tool on the saved JSON to consume it programmatically.
# Examples — every command accepts --raw for byte-stable v5.0.0 output
RAW_FLAG=""
if echo "$ARGUMENTS" | grep -q -- "--raw"; then RAW_FLAG="--raw"; fi
# /config-audit posture --raw
# /config-audit tokens --raw
# /config-audit fix --raw
All Commands
Core
| Command | Description |
|---|---|
/config-audit |
Full audit with auto-scope detection |
/config-audit posture |
Quick scorecard with A-F grades per area (10 areas) |
/config-audit tokens |
prompt-cache token hotspots; optional --accurate-tokens API calibration |
/config-audit manifest |
Ranked table of every system-prompt token source |
/config-audit feature-gap |
Deep analysis of features you're not using |
/config-audit optimize |
Optimization lens — config that works but fits a better mechanism (procedure→skill, lifecycle→hook, path→rule, never→permission) |
/config-audit fix |
Auto-fix deterministic issues; a copy of every changed file is saved first so you can roll back with one command |
/config-audit rollback |
Restore configuration from a saved copy |
Planning & Implementation
| Command | Description |
|---|---|
/config-audit plan |
Generate prioritized action plan from audit findings |
/config-audit implement |
Execute action plan; a copy of every changed file is saved first, and a verification pass runs after |
/config-audit interview |
Set preferences to customize the action plan (optional) |
Monitoring
| Command | Description |
|---|---|
/config-audit drift |
Compare current config against a saved baseline |
/config-audit plugin-health |
Audit plugin structure and the metadata block at the top of each command/agent file |
/config-audit whats-active |
Show active plugins/skills/MCP/hooks/CLAUDE.md with token estimates |
Utility
| Command | Description |
|---|---|
/config-audit status |
Show current session state and progress |
/config-audit cleanup |
Clean up old session directories |
Advanced (workflow phases)
| Command | Description |
|---|---|
/config-audit discover |
Run only the discovery phase (find config files) |
/config-audit analyze |
Run only the analysis phase (generate report) |
Plain-language vocabulary
The toolchain uses these terms when describing findings:
| User-facing label | What it means |
|---|---|
| Fix this now | Something is broken or risky and should be addressed immediately |
| Fix soon | High-priority issue worth scheduling this week |
| Fix when convenient | Real issue but not urgent |
| Optional cleanup | Tidy-up that improves polish but isn't required |
| FYI | Informational; no action expected |
| Configuration mistake | A configuration file has an error or omission |
| Conflict | Two configuration sources disagree |
| Wasted tokens | Configuration is loading content that costs tokens without payback |
| Missed opportunity | A Claude Code feature you aren't using that could help your project |
| Dead config | Configuration that has no effect (e.g., a permission that's also denied) |
Use --raw if you'd rather see the v5.0.0 verbatim output (technical IDs and raw severity).
Scope Override
By default, /config-audit auto-detects scope from your current directory:
- Inside a git repo → scans the repo
- In
$HOME→ scans global config only - Elsewhere → scans current directory
Override with: /config-audit current, /config-audit repo, /config-audit home, /config-audit full
Typical Workflows
First time? Just run /config-audit.
Want to fix things? Run /config-audit then /config-audit fix.
Full optimization:
/config-audit— see what you have/config-audit plan— create action plan/config-audit implement— execute with backups
Track changes over time:
/config-audit drift --save— save baseline- (make changes)
/config-audit drift— see what changed