ai-agent-memory: Skill for Claude Code

Review Sessions

When the user says "review sessions", "session review", or invokes /review-sessions, analyse coding-agent session data to identify workflow improvements, prompt-quality patterns, error trends, and security hygiene issues.

Works across both supported agents:

Agent	Where sessions live	Format
GitHub Copilot CLI	~/.copilot/session-state/{session-id}/events.jsonl (+ plan.md, checkpoints/, command-history-state.json)	JSONL event stream — typed events (session.start, session.mode_changed, turn.user, turn.assistant, tool.invoked, …) with timestamp, id, parentId
Claude Code	~/.claude/projects/{project-hash}/{session-id}.jsonl (+ ~/.claude/history.jsonl)	JSONL conversation turns — `type: user

Resolve the memory root

1. If memory/schema.md exists in the current workspace, set WIKI_ROOT to memory/.
2. Else if schema.md exists, set WIKI_ROOT to the current directory.
3. Else follow the memory-wiki path in AGENT.md.
4. If no folder containing both schema.md and index.md can be resolved, report the missing path and stop.

All ops/... paths below are relative to WIKI_ROOT.

Scope modifiers

Interpret these optional modifiers from the user's request:

• --agent {copilot,claude,all} — source to review; default all
• --all — ignore the watermark and review every session
• --project SLUG — match the session cwd or git root
• --since YYYY-MM-DD — include sessions on or after this date
• No modifiers — incremental review using ops/review-state.json

Step 1: Discover and normalize sessions directly

Use the agent's native file listing, search, JSON/JSONL reading, and reasoning capabilities. Do not require or create a helper extraction script.

1. Inventory session files for the selected agent source(s).
2. For incremental mode, read ops/review-state.json if it exists and exclude namespaced file IDs already present in watermark.reviewed_files. A file ID is copilot:{absolute-path} or claude:{absolute-path}.
3. Apply project and date filters before reading large files.
4. Read each JSONL file in manageable chunks. Parse each line independently; count and report malformed lines instead of silently discarding them.
5. Normalize records into this common in-memory model:
   • agent: copilot or claude
   • session_id, timestamp
   • kind: user_turn, assistant_turn, tool_call, tool_result, error, or mode_change
   • text, tool_name, success, duration_ms when available
6. Copilot mapping: use event type values such as turn.user, turn.assistant, tool.invoked, tool completion/error events, and session.mode_changed.
7. Claude mapping: use type: user|assistant records and nested tool_use / tool_result content; infer plan-mode changes only when explicitly represented.
8. Keep the processed namespaced file IDs for the state update in Step 6.