general-counsel-advisor
Pass
Audited by Gen Agent Trust Hub on Oct 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external JSON files, which is then rendered into the agent's context as Markdown reports. This creates a potential surface for an attacker to influence agent behavior by providing malicious input data.
- Ingestion points: The skill reads from
legal_risk_inputs.json,contracts.json, andregulatory_inputs.jsonvia the workflows defined inSKILL.md. - Boundary markers: Absent. The Markdown generated by the Python scripts in the
scripts/directory is interpolated into the prompt without protective delimiters or instructions to ignore embedded commands. - Capability inventory: The skill possesses the capability to execute local Python scripts and read from the filesystem, including accessing the user's
~/.claude/company-context.mdfile. - Sanitization: Absent. The processing scripts (
legal_risk_register.py,contract_portfolio_analyzer.py,regulatory_calendar_generator.py) do not sanitize or escape JSON string values before they are rendered into the Markdown output.
Audit Metadata