revid
Pass
Audited by Gen Agent Trust Hub on Aug 25, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection via user-supplied arguments and file paths.
- Ingestion points: Command-line arguments and file paths passed to
scripts/revid_api.sh. - Boundary markers: Absent; there are no instructions for the agent to validate or delimit untrusted inputs before execution.
- Capability inventory: The skill can perform network requests (
curl) and read local file contents. - Sanitization: Partial; while the script uses
--data-urlencodefor search queries, it performs raw string interpolation for other parameters like project IDs and file paths, which could lead to JSON or parameter injection. - [DATA_EXFILTRATION]: The
api_post_jsonfunction inscripts/revid_api.shuses thecurl --data @option, which allows the script to read and transmit the contents of any file path provided as an argument. This could be exploited to exfiltrate sensitive local data (such as SSH keys or environment files) if the agent is manipulated into passing unauthorized file paths to the API. - [COMMAND_EXECUTION]: The skill executes shell commands using
curlvia a bundled Bash script to interact with the Revid.ai platform.
Audit Metadata