pm-research-methodologies
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The research methodology requires the agent to ingest and process data from various untrusted external sources.\n
- Ingestion points: External data is ingested through search queries and the browsing of primary records, press, and social media platforms like X and Telegram (SKILL.md, Stage 2).\n
- Boundary markers: The instructions do not mandate the use of delimiters or security warnings when the agent retrieves external content, potentially allowing embedded instructions in that data to influence the agent.\n
- Capability inventory: The agent is authorized to perform web searches, navigate to external URLs, and record synthesis to an output destination (SKILL.md, Stages 2 and 4).\n
- Sanitization: There are no instructions for sanitizing or escaping retrieved content before it is used for synthesis or next steps.\n- [NO_CODE]: The skill is composed strictly of instructional markdown files and does not distribute any executable scripts, binaries, or automated configuration files.
Audit Metadata