The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data including error logs, stack traces, and code reports provided in the execution context (SKILL.md). While it mandates the redaction of secrets, it lacks explicit boundary markers or delimiters for separating this external data from internal instructions. Given the skill's capabilities to execute shell commands and modify code (SKILL.md), malicious instructions embedded in a processed log or trace could potentially influence the agent's actions.
[DYNAMIC_EXECUTION]: The instructions guide the agent to dynamically create and run CLI scripts, HTTP tests, and throwaway harnesses (SKILL.md) to reproduce bugs. This runtime code generation and execution is a core functional requirement of the skill but constitutes a dynamic execution pattern that requires oversight.
[SAFE]: The skill includes security-conscious instructions that require the agent to redact secrets from all commands, logs, and reports, and specifically mandates keeping credentials in environment variables (SKILL.md). This aligns with industry standards for secure credential management.

systematic-debugging