Agent Continuity
Pass
Audited by Gen Agent Trust Hub on Aug 15, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill relies on reading untrusted data from files like
.agent/SESSION.mdandtasks/todo.mdto determine next steps and sanity checks. This creates a vulnerability surface where an attacker with write access to the repository could influence agent behavior. - Ingestion points:
.agent/SESSION.md,.agent/PROJECT.md, andtasks/todo.md. - Boundary markers: No explicit delimiters or instructions to ignore embedded commands are present in the skill definition.
- Capability inventory: The skill expects the agent to execute shell commands (e.g., sanity builds/tests) and perform file system operations based on the state found in these files.
- Sanitization: There is no requirement for the agent to sanitize or validate the content of the session files before processing.
Audit Metadata