The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill's primary function involves analyzing untrusted external source code and binary assets. This creates an inherent surface for indirect prompt injection where malicious instructions embedded in the target code could attempt to influence agent behavior during the audit.
Ingestion points: SKILL.md instructs the agent to perform static and binary inspection of target codebases.
Boundary markers: The skill does not define specific delimiters or instructions to ignore embedded commands within the analyzed data.
Capability inventory: The workflow involves file system inspection, binary disassembly, and potentially executing build/test commands as part of verification.
Sanitization: No sanitization methods for the ingested content are mentioned.
[COMMAND_EXECUTION]: The workflow in SKILL.md suggests the execution of verification commands and test suites (e.g., npm test) to validate remediations. These are standard procedures for security auditing and remediation verification.
[NO_CODE]: The skill consists entirely of instructional markdown and YAML configuration metadata. It does not include any standalone executable scripts or binaries.

offensive-deserialization