report
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill performs local data aggregation and formatting for security reporting purposes. It reads from defined project directories (.appsec/) and writes results to the same area. No network activity, remote code execution, or credential access was identified.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests external data to generate reports, though the risk is mitigated by the environment configuration. 1. Ingestion points: SKILL.md Step 1 (Gather Data) reads findings from .appsec/findings.json, .appsec/fixed-history.json, and raw scanner output in .appsec/scanners/. 2. Boundary markers: No specific delimiters or warnings for embedded instructions are specified for the output report. 3. Capability inventory: The skill is restricted by 'disable-model-invocation: true', preventing the model from being manipulated during execution. It has file-read and file-write access but no network or subprocess capabilities. 4. Sanitization: No explicit sanitization or filtering of findings data is described in the workflow.
Audit Metadata