dcf-model
Warn
Audited by Gen Agent Trust Hub on Sep 5, 2026
Risk Level: MEDIUMMETADATA_POISONINGINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [METADATA_POISONING]: The 'author' field in the frontmatter claims authorship by 'Anthropic (adapted by Nous Research)', contradicting the provided context of 'zapabob'. This misleading attribution poses a concern regarding the skill's origin.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the web and MCP servers to build financial models.
- Ingestion points: Fetches market data and financials from sources like SEC EDGAR and MCP servers (SKILL.md).
- Boundary markers: Mandates step-by-step user verification and detailed cell comments for source attribution.
- Capability inventory: Capability to execute local Python scripts (recalc.py, validate_dcf.py) and write Excel files.
- Sanitization: Lacks explicit instructions for sanitizing external strings before insertion into the model.
- [COMMAND_EXECUTION]: The intended workflow requires the agent to run local Python scripts ('recalc.py' and 'validate_dcf.py') for model validation.
- [EXTERNAL_DOWNLOADS]: The skill uses the 'requests' library to fetch data from well-known financial services and government databases.
Audit Metadata