ib-deck
Pass
Audited by Gen Agent Trust Hub on Jun 17, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a local Python script
infra/deck_renderer.pyto convert generated HTML into PDF format. TheTICKERargument provided by the user is used to construct the input and output file paths. While this presents a potential command injection surface, the risk is mitigated in practice by the agent's typical use of data discovery tools to validate tickers before execution. - [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) because it processes untrusted data from SEC filings and user-provided firm names.
- Ingestion points: User-supplied
TICKERand firm name; external SEC filings and Daloopa financial records retrieved via MCP. - Boundary markers: The instructions do not specify any delimiters or warnings to ignore instructions embedded within the retrieved data.
- Capability inventory: The skill can write files to the
reports/directory and execute thedeck_renderer.pyscript via subprocess. - Sanitization: There is no evidence of sanitization or validation of the retrieved text content before it is processed into the HTML deck.
- [DATA_EXFILTRATION]: The skill connects to
daloopa.comto fetch financial data and cite sources. As this is the author's own domain and core to the skill's stated purpose, this activity is considered standard and safe functionality.
Audit Metadata