The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Bash tool to execute local utility scripts and hooks, such as python .claude/scripts/analyze_errors.py and .claude/hooks/posttool/error-collector.sh. These operations are essential for its purpose of monitoring and analyzing the agent's internal tool usage and failure patterns.
[PROMPT_INJECTION]: The skill presents a surface for indirect prompt injection because it processes untrusted data from tool outputs (error messages) to auto-generate rules.
Ingestion points: Error messages and tool inputs are captured and stored in .claude/logs/errors.jsonl.
Boundary markers: The instructions do not define specific delimiters or "ignore instructions" guards for the data being analyzed.
Capability inventory: The skill utilizes Bash, python, Read, and Grep tools.
Sanitization: No explicit sanitization or filtering of the ingested error messages is documented in the provided logic.

errors