The Agent Skills Directory

[SAFE]: The skill implements a "diagnose-first" protocol that prevents automatic code modification and requires explicit human authorization for all patches, as documented in SKILL.md and references/debug-policy.md.
[SAFE]: The Python script scripts/safe_debug.py facilitates error diagnosis by reading provided error text or files. While it accesses file content, it does so within the context of a debugging tool and lacks network or code execution capabilities that would enable data exfiltration.
[SAFE]: The skill handles untrusted data (error tracebacks) with appropriate safeguards, such as truncating content excerpts and utilizing markdown code block boundaries in its generated reports, effectively mitigating the risk of indirect prompt injection.

safe-debug