lode-decision-roadmap
Pass
Audited by Gen Agent Trust Hub on May 12, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill invokes local Python scripts (
lode_raw.py) via the shell to resolve configuration paths and project metadata. It passes the current working directory ($PWD) as an argument, which could be exploited if not handled securely by the underlying platform. - [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it ingests and processes untrusted data from JSON files within the knowledge vault to generate narrative summaries.
- Ingestion points: Data is read from files at
{vault}/raw/weeks/{YYYY-WNN}/{slug}.jsonand{vault}/raw/artifacts/{slug}.json. - Boundary markers: No explicit boundary markers or instructions to ignore embedded malicious prompts are present in the workflow.
- Capability inventory: The agent has the ability to execute shell commands (via the Python helper script) and write to the local filesystem.
- Sanitization: There is no evidence of sanitization or filtering of the
summary,context, or other decision fields before they are used for synthesis.
Audit Metadata