The Agent Skills Directory

[SAFE]: No malicious patterns, obfuscation, or unauthorized data access were detected. The skill is purely methodological and serves as a guide for systematic debugging.
[INDIRECT_PROMPT_INJECTION]: The skill involves processing external, potentially untrusted data such as bug reports and application logs.
Ingestion points: User-provided bug reports, stack traces, and application logs referenced in SKILL.md and template.md.
Boundary markers: The skill encourages the use of structured Markdown headers (e.g., ## Data Gathered) in its templates, which act as delimiters between data and instructions.
Capability inventory: The skill utilization of Bash, Read, Write, and Edit tools to perform experiments and verify fixes.
Sanitization: No specific input sanitization or filtering logic is defined within the skill instructions, relying on the agent's internal safety guardrails.

systematic-debugging