The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted source code files, creating a surface for indirect prompt injection if malicious instructions are embedded in the analyzed code.\n- Ingestion points: Source code files targeted by the audit, such as upload handlers and file processing logic identified in SKILL.md.\n- Boundary markers: The skill lacks explicit instructions or markers to isolate the analyzed code content from the agent's control logic.\n- Capability inventory: The agent can execute external security scanners (semgrep, bandit, brakeman) and perform manual analysis of the code.\n- Sanitization: There is no evidence of content sanitization or escaping applied to the analyzed files before they are processed by the agent.\n- [EXTERNAL_DOWNLOADS]: The workflow involves running semgrep with the auto configuration, which fetches scanning rules from the official Semgrep registry. This is an industry-standard practice for security auditing and involves a well-known service.\n- [SAFE]: The provided detection patterns in references/detection-patterns.md are educational and accurately represent common security vulnerabilities and their mitigations.\n- [SAFE]: No indicators of malicious intent, such as data exfiltration, obfuscation, or persistence, were identified in the skill instructions or reference files.

file-upload