paper-context-resolver
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external data from research papers and software repositories to assist in reproduction tasks.
- Ingestion points: Data enters the agent context through "target repo metadata", "reproduction-critical question", and "existing README or repo evidence" as specified in the input expectations of
SKILL.md. - Boundary markers: The instructions in
references/paper-assisted-reproduction.mdinclude a "Conflict rule" that requires the agent to record discrepancies between sources and maintain a "README-first" policy, providing logical delimiters for source processing. - Capability inventory: The provided files do not define specific automated tool usage, shell execution, or file-writing capabilities; the skill focuses on instructional logic for the agent's analytical reasoning.
- Sanitization: There are no explicit instructions for sanitizing, escaping, or validating the content retrieved from external repository files or research papers.
Audit Metadata