The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data such as logs, stack traces, and test outputs to guide the debugging process.
Ingestion points: The skill processes external data during the 'Hypothesize' (Step 3) and 'Experiment' (Step 4) phases described in SKILL.md.
Boundary markers: There are no specific delimiters or instructions for the agent to disregard potential prompts embedded within the debug data.
Capability inventory: The skill utilizes file editing (Edit) and command execution tools (e.g., python3, pytest, grep) as detailed in SKILL.md.
Sanitization: No explicit sanitization or validation of the ingested data is mentioned.
[DYNAMIC_EXECUTION]: The skill encourages creating reproduction scripts or using python3 -c for testing in SKILL.md, which involves executing code generated based on the debugging context.
[COMMAND_EXECUTION]: The skill makes use of shell commands and test runners like pytest, python3, and search tools like grep to locate and verify defects, which are standard developer operations but represent a capability tier for the agent.

cc-debugging