handoff-resume
Pass
Audited by Gen Agent Trust Hub on Aug 17, 2026
Risk Level: SAFEPROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection through the content of the handoff document it processes.
- Ingestion points: The skill reads a file path provided via
$ARGUMENTS[0]inSKILL.md. - Boundary markers: Absent. There are no instructions to delimit the external content or warn the agent to ignore embedded instructions.
- Capability inventory: The skill has the ability to read files and list directory contents.
- Sanitization: Absent. The skill specifically instructs the agent to trust the "Work Done" section of the file as an authoritative record, increasing the likelihood that malicious instructions within the file will be followed.
- [DATA_EXFILTRATION]: The skill accepts a file path as a direct input argument and reads it into the agent's context. This allows for the exposure of sensitive local files (such as
.env, SSH configurations, or cloud credentials) if an unintended path is provided to the agent.
Audit Metadata