gong
Pass
Audited by Gen Agent Trust Hub on Mar 4, 2026
Risk Level: SAFECREDENTIALS_UNSAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [CREDENTIALS_UNSAFE]: The skill manages authentication by reading from a local JSON file at
~/.config/gong/credentials.json. While necessary for API access, this places sensitive credentials on the local filesystem. - [COMMAND_EXECUTION]: In
scripts/gong.sh, arguments such as call IDs are interpolated into shell command strings without sanitization. This could allow for shell command expansion if the input contains shell metacharacters like backticks or$(). - [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection when processing transcripts or call data from the Gong API. Maliciously crafted content in a call transcript could influence the agent's behavior.
- Ingestion points:
scripts/gong.sh(transcript and call commands) andSKILL.mdexamples. - Boundary markers: Absent; the transcript text is passed directly into the agent's context without delimiters or warnings.
- Capability inventory: The skill can read local files, execute shell commands via
curl, and perform network operations. - Sanitization: No sanitization or content filtering is applied to the API responses before they are outputted.
Audit Metadata