The Agent Skills Directory

[SAFE]: The skill provides a comprehensive methodology for post-breach remediation, covering root cause analysis, control gap closure, and regulatory accountability. It integrates industry-standard references such as GDPR Art. 33 and ISO/IEC 27001, and mentions legitimate security tools like KnowBe4 for training simulation.\n- [DYNAMIC_EXECUTION]: The included Python script scripts/process.py is a deterministic utility for generating remediation plans and tracking metrics. It relies on the Python standard library for basic data processing and contains no unsafe dynamic execution sinks such as eval() or exec().\n- [DATA_EXFILTRATION]: No network operations, data exfiltration patterns, or unauthorized file access attempts were detected. The skill's scripts perform local data transformation and output findings to standard output.\n- [INDIRECT_PROMPT_INJECTION]: The skill provides a surface for processing breach details into structured plans. (1) Ingestion points: breach_reference and root_causes parameters in scripts/process.py. (2) Boundary markers: The script employs a static dictionary lookup to map input strings to predefined remediation actions. (3) Capability inventory: No dangerous system-level capabilities (e.g., file writing, network requests, or shell command execution) are exposed. (4) Sanitization: Input is validated against a fixed set of template keys.\n- [OBFUSCATION]: All analyzed markdown and code files are written in clear text. No Base64 encoding, zero-width character embedding, or homoglyph attacks were identified.

breach-remediation