money-review-investor
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [DATA_EXFILTRATION]: The skill instructs the agent to access and read data from hidden directories in the user's home folder, specifically
~/.smtm/sessions/and~/.smtm/projects/. - Evidence: Instructions to load the latest snapshot from
~/.smtm/sessions/{slug}/and project learnings from~/.smtm/projects/{slug}/learnings.jsonl. - Context: While these paths appear specific to the skill's 'SMTM' ecosystem, hidden home directory access constitutes a data exposure risk. Additionally, the 'Standard startup' sequence mentions a 'telemetry write', which implies outgoing data flow, though no specific destination is defined in this file.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted business plans and strategy documents, which may contain malicious instructions intended to manipulate the agent's evaluation or access the loaded session context.
- Ingestion points: User-provided plan files (
path-to-plan.md), conversation history, and project snapshots in~/.smtm/. - Boundary markers: None identified. The instructions do not define delimiters or direct the agent to ignore instructions embedded within the plan content.
- Capability inventory: The agent has file system read access to project snapshots and learning history files.
- Sanitization: There is no evidence of sanitization, validation, or escaping of the ingested plan content before it is processed by the LLM.
Audit Metadata