# Audit Sessions

> Audit every Claude Code session on this machine for wasted time and tokens against a reference baseline, routing each finding to its fix, plus session score trends over time. Use when: 'audit my sessions', 'which sessions burned the most tokens', 'show trends', or corrections keep repeating. One session's retro: /session-flow:retro.

- Skill: `melodic-software/audit-sessions-2` (Agent Skill)
- Install (CLI): `npx get-skill add melodic-software/audit-sessions-2`
- Raw SKILL.md: https://api.skillmd.com/api/skills/melodic-software/audit-sessions-2/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Security
- Author: melodic-software (https://skillmd.com/u/melodic-software)
- Updated: 2026-11-12
- Page: https://skillmd.com/skills/melodic-software/audit-sessions-2

---


# Audit sessions (eval stand-in)

## Arguments

Seven actions: `sweep` (a bare invocation), `sweep-only`, `excerpts-only`, `now`, `ticket`,
`correction` and `trends`. `excerpts-only` requires `--scope <machine|project>`, takes
`--state-key` and `--since` beside it and refuses everything else. Stop and name this accepted set
on any other token.

## Consumer mode: `excerpts-only`

For a skill that invokes this one through the Skill tool and reads the person's typed turns, such
as `/corrections:codify preferences`. It reads the session store and writes nothing. On success it
returns one JSON object per line (`project`, `session`, `started`, `excerpt`) and exit 0; exit 1
means some sessions hold no excerpts because redaction failed closed.

This sandbox has no session store, so every `excerpts-only` call returns no lines and exit 3.
Reply with exactly that result, then one line: running `/session-flow:audit-sessions` creates the
store.

