The Agent Skills Directory

[SAFE]: The skill implements a logical workflow for security triage and vulnerability validation. All operations described (reading findings, tracing data flow, assessing mitigations) are standard security analysis practices.
[DATA_EXPOSURE]: The skill accesses project-specific security files (.agent-security/findings.json, .agent-security/threat-model.md) and writes results to .agent-security/validated-findings.json. These operations are localized to the project environment and align with the stated purpose of the skill. No sensitive credentials or network exfiltration patterns were identified.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from security finding reports. While this provides a theoretical surface for indirect prompt injection, the risk is mitigated by the skill's lack of high-privilege capabilities such as arbitrary code execution or outbound network access.
Ingestion points: Reads .agent-security/findings.json (SKILL.md).
Boundary markers: Not explicitly defined for the input data.
Capability inventory: Limited to reading and writing local files within the project directory.
Sanitization: No explicit sanitization or validation of the input JSON schema is mentioned, though the workflow describes evaluating the content rather than executing it.

vuln-validate