The Agent Skills Directory

[PROMPT_INJECTION]: The skill inherently processes untrusted data (raw bug bounty findings and proof-of-concept payloads) as part of its primary report-generation function, which creates a surface for indirect prompt injection. However, given its intended use case for security professionals, this is a functional requirement. • Ingestion points: Raw notes, requests, and PoCs provided by the user in SKILL.md. • Boundary markers: Not explicitly defined in the instructions for user-supplied input. • Capability inventory: The skill performs text transformation and formatting; no active scripts for network or file system operations were found in the provided files. • Sanitization: No explicit instructions for sanitizing or escaping the untrusted input content are present.
[COMMAND_EXECUTION]: SKILL.md contains an instruction to run a local script 'scripts/cvss-scorer.py' for calculating vulnerability severity. While this script file was not included in the analysis bundle, the reference is standard for bug bounty tooling and does not indicate malicious behavior.
[SAFE]: The reference files contain numerous examples of malicious payloads (e.g., SQL injection, XSS with Base64 encoding, XXE, and SSRF bypasses). These are documented strictly as vulnerability examples for reporting purposes and are not executed by the skill itself.

bugbounty-reporter