Slash Command

/adlc-maintain

From adlc

Run the decay-driven ADLC maintenance checks — stale skills, hot files to re-prosecute, and gate calibration (C10/C12).

Popularity

Stars

Forks

Invocation

How this command is triggered — by the user, by Claude, or both

Slash command

/adlc:adlc-maintain (no arguments)

Model invocable

No pre-commands

Referenced Files

Files this command reads when invoked

adlc/cli

Context Preview

The summary Claude sees in its command listing — used to decide when to auto-load this command

# /adlc-maintain — fight decay (C10 / C12 + calibration)

Some assumptions rot over time or after a model/repo change: skill cache
metadata goes stale, files churn enough to deserve re-prosecution, and gates that
once held may now be defeatable. This command runs those checks. It is idle-time
work — run it on a schedule or after a model upgrade.

Prerequisite: `adlc --version` works (else `npm i -g @adlc/cli`). Report a single
honest summary at the end, including any check that did not apply.

## 1. Skill rot — stale validation metadata (C10)



- Exit `0`: skills are fresh.
- Exit `2`: one...

Command Content

67 lines · ~705 tokens

Stats

LanguageJavaScript

Stars3

Forks1

MaintenanceExcellent

Last CommitJun 18, 2026

Actions

View Source View Plugin View on GitHub View README

Stats

Actions

/adlc-maintain — fight decay (C10 / C12 + calibration)

Some assumptions rot over time or after a model/repo change: skill cache metadata goes stale, files churn enough to deserve re-prosecution, and gates that once held may now be defeatable. This command runs those checks. It is idle-time work — run it on a schedule or after a model upgrade.

Prerequisite: adlc --version works (else npm i -g @adlc/cli). Report a single honest summary at the end, including any check that did not apply.

1. Skill rot — stale validation metadata (C10)

adlc skill-rot <skill-dir-or-glob> --json

Exit 0: skills are fresh.
Exit 2: one or more skills have stale validation metadata — list them and recommend re-validating (and --write to stamp freshness once re-checked).
Exit 1 with nothing to verify: the targeted skills carry no validation metadata to check — this is informational, not a failure. Note it.

2. Model ratchet — hot files to re-prosecute (C12)

adlc model-ratchet --dry-run --json

Lists the highest-churn / highest-dependency files (score), which are the best candidates to re-prosecute after model or repo drift. This is a plan, not a gate — it does not fail. Report the top files and suggest running the prosecutor (/adlc-prosecute or the prosecutor subagent) against them.
With a --review-cmd, model-ratchet can run a review over those files and append findings to .adlc/findings.jsonl (which later feeds /adlc-distill).

3. Gate fuzzing — can hostile candidates defeat the gates? (calibration)

Only run this if a gate suite exists at .adlc/gate-suite.json; without one the tool exits 1 (Gate suite not found) — note that calibration was skipped.

adlc gate-fuzzing --suite .adlc/gate-suite.json --prompt-only

LLM-backed: in Claude, use --prompt-only and play the adversary yourself against each gate in the suite. A gate you can defeat is a calibration gap — report it; the gate needs strengthening (this is the gate-fuzzing exit-2 condition when run with a provider).

4. Summarize

Report: stale skills (if any), the top hot files to re-prosecute, gate-fuzzing result or why it was skipped, and the recommended next actions. Repeated findings surfaced here flow into /adlc-distill.

Scheduling

The deterministic checks here (skill-rot, model-ratchet) are keyless and run well on a cron — see the ready-to-use workflow at docs/ci/adlc-maintenance.yml. The LLM-backed gate-fuzzing runs via a scheduled Claude routine (/schedule invoking /adlc-maintain), where Claude is the model and no API keys are needed.