omarchy-plugin-publish
Pass
Audited by Gen Agent Trust Hub on Nov 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The
scripts/prepare_submission.pyscript executes thegit remote get-url origincommand viasubprocess.runto resolve the repository URL from the local environment. - [COMMAND_EXECUTION]: The skill instructions in
SKILL.mdandreferences/marketplace-submission.mdguide the agent to perform administrative actions using thegh issue createcommand, relying on the host's authenticated session. - [INDIRECT_PROMPT_INJECTION]: The
prepare_submission.pyscript ingests thenameandidfields from a user-providedmanifest.jsonand interpolates them directly into a markdown template. If these fields contain malicious instructions, they could influence the agent's behavior or the content of the resulting GitHub issue. - Ingestion points: The
manifest.jsonfile in the target plugin directory, read viaPath.read_text(). - Boundary markers: The script uses markdown headers (e.g.,
### Category) to structure the output, but does not implement explicit delimiters or warnings to ignore instructions within the manifest data. - Capability inventory: The skill is capable of running local Python scripts, reading and writing files in the plugin directory, and executing shell commands (
git,gh). - Sanitization: The script validates that manifest fields are non-empty strings but does not escape or filter their content before interpolation into the markdown draft.
Audit Metadata