package-once-green
Pass
Audited by Gen Agent Trust Hub on Oct 3, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSDYNAMIC_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The
greenlauncher script fetches dependencies from the vendor's official repositories on GitHub (github.com/getcolors/once.gitandgithub.com/getcolors/green.git) using specific commit SHAs for version pinning. - [DYNAMIC_EXECUTION]: The launcher uses Babashka to dynamically load and resolve Clojure namespaces from the fetched libraries at runtime to execute deployment workflows.
- [COMMAND_EXECUTION]: The skill instructions involve running the
./greenCLI tool to perform various deployment operations including provisioning, builds, and dry-runs. - [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided configuration in
colors.ymland environment variables to perform deployment tasks. - Ingestion points: Ingests configuration from
colors.ymland environment variables (documented inSKILL.mdandgreen). - Boundary markers: Explicit instructions are provided to the agent to avoid reading private files (such as
.envrc.private) and to only request variable names rather than their secret values. - Capability inventory: The skill manages SSH keys, performs network API calls to cloud providers (DigitalOcean, Cloudflare, Resend, GitHub), writes to a
.colors/directory, and executes shell commands via thegreenlauncher. - Sanitization: The skill utilizes validation steps through
./green buildand dry-run execution prior to convergence.
Audit Metadata