trial-budget-brief
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a bundled Python script (
scripts/summarize_trial_cost.py) to calculate trial costs. This is a controlled execution of a local file provided within the skill's own package to ensure deterministic calculations. - [INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted external data related to hiring plans and budgets provided by the user.
- Ingestion points: User-provided trial descriptions, cost estimates, and role details are processed in
SKILL.md(Step 5) to generate cost summaries. - Boundary markers: The instructions guide the agent to use specific JSON structures and reference documentation (
references/brief-method.md) to delimit and validate input. - Capability inventory: The skill uses the script to calculate totals and comparison data, which is then formatted into a report according to the format in
references/report-format.md. - Sanitization: The skill utilizes a shared helper module
scripts/_common.pyto provide robust type checking and validation for all input fields (strings, numbers, dates), including regex patterns to detect potential personal information (PII) such as emails and phone numbers before processing.
Audit Metadata