concept-graph
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill has an ingestion surface as it processes
output/TUTORIAL_SPEC.mdwithinscripts/run.py. However, the risk is negligible as the data is transformed into a structured YAML graph (outline/concept_graph.yml) for planning purposes, and the skill lacks dangerous capabilities such as network access or arbitrary command execution. - [COMMAND_EXECUTION]: The skill executes a local Python script
scripts/run.py. The script uses standard argument parsing and performs file I/O within the defined workspace, posing no threat to the host system.
Audit Metadata