gh-upload
Pass
Audited by Gen Agent Trust Hub on Aug 18, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes shell commands through
gh,curl, andffmpegto retrieve repository metadata, upload binary data, and transcode video files. - [DATA_EXFILTRATION]: Utilizes
curlto transmit local files tohttps://uploads.github.com. This network operation targets a well-known service as part of the skill's primary functionality for embedding assets into GitHub PRs and issues. The skill also accesses authentication tokens viagh auth tokento authorize these requests. - [INDIRECT_PROMPT_INJECTION]: The skill processes external data which creates an injection surface:
- Ingestion points: User-provided file paths (
$FILE), display names ($NAME), MIME types ($MIME), and repository identifiers interpolated into commands inSKILL.md. - Boundary markers: None provided in the command templates.
- Capability inventory: Subprocess execution via
Bashforcurl,gh, andffmpegtools. - Sanitization: None; the skill instructions suggest manual percent-encoding but provide no automated validation.
Audit Metadata