The Agent Skills Directory

[COMMAND_EXECUTION]: The skill relies on shell command execution for its core functionality. It executes scripts/analyze-session-logs.sh, which uses jq, grep, awk, and sed to parse JSONL logs. It performs git operations such as git log and git diff to analyze repository changes. It also references and invokes external scripts from the extracting-session-data skill, including list-sessions.sh, extract-data.sh, and filter-sessions.sh.
[PROMPT_INJECTION]: The skill presents an indirect prompt injection surface because it ingests and analyzes conversation history from session logs to generate recommendations. 1. Ingestion points: Reads JSONL logs from ~/.claude/projects/ and collects feedback from participating sub-agents. 2. Boundary markers: The instructions do not specify explicit delimiters or boundary markers when presenting analyzed log data to the agent. 3. Capability inventory: The skill can modify core configuration files like .claude/CLAUDE.md and SKILL.md via the Edit tool based on its analysis. 4. Sanitization: No explicit sanitization or filtering of processed log content is described before generating configuration update suggestions.

retrospecting