grow-app
Warn
Audited by Gen Agent Trust Hub on Aug 23, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructs the user to download and install external code via the command
npx skills add wondelai/skills/<slug> --global. The source repository 'wondelai/skills' is not associated with the skill's official author (manish31-oss) and does not belong to a recognized trusted vendor. - [COMMAND_EXECUTION]: The skill contains a fallback rule that prompts the user to execute shell commands (
npx) to add missing dependencies. While this requires user interaction, it facilitates the execution of external binaries and packages within the user's environment. - [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection through its ingestion of local files. It reads state and previous decisions from
docs/GROW-APP-PLAN.mdand other artifacts to resume its journey, which can influence the agent's logic if those files contain embedded instructions. - Ingestion points: Reads
docs/GROW-APP-PLAN.md,docs/PRODUCT.md,docs/METRICS.md, and other files in thedocs/folder (Operating Rules 1 and 7). - Boundary markers: Absent. The skill does not use delimiters or instructions to ignore potential commands embedded in the processed documentation.
- Capability inventory: The skill has the ability to write to the local filesystem and provides commands for shell execution.
- Sanitization: Absent. The agent summarizes the journey state directly from the read files without validation or filtering of the content.
Audit Metadata