jobs-to-be-done-skill
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill instructs the agent to write session metadata and execution metrics to a local JSON file (~/.claude/skill-analytics/last-outcome-jobs-to-be-done-skill.json). This is a standard telemetry feature for monitoring skill performance and does not involve the exfiltration of sensitive user data or unauthorized network access.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user input describing companies or markets for analysis, which is then processed and recorded in a local log file.
- Ingestion points: User-provided text describing the [company/product/market] analysis target.
- Boundary markers: The instructions lack explicit delimiters to separate user-supplied content from analytical directives.
- Capability inventory: The skill uses a local file-write capability to update its telemetry logs.
- Sanitization: There are no instructions for sanitizing or escaping user input before it is recorded in the analytics JSON.
Audit Metadata