The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/run-lifecyclemodel-automated-builder.mjs acts as a Node.js wrapper for the tiangong CLI tool. It interprets user-supplied flags like --input and --out-dir to execute canonical CLI commands for building, validating, and publishing lifecycle model artifacts. Evidence of this is found in the runBuild and runDelegatedLifecyclemodelCommand functions which orchestrate the tiangong lifecyclemodel invocations.
[PROMPT_INJECTION]: The skill presents a surface for indirect prompt injection because it ingests and processes user-defined JSON manifests to determine the parameters of its file system and CLI operations. 1. Ingestion points: Manifest files read via the --input, --manifest, or --request parameters in scripts/run-lifecyclemodel-automated-builder.mjs. 2. Boundary markers: Absent. There are no explicit delimiters or 'ignore' instructions provided to the agent to mitigate instructions potentially embedded in the manifest data. 3. Capability inventory: The skill performs shell command execution through the tiangong CLI and writes various JSON and Markdown reports to the local file system as detailed in the Workflow section of SKILL.md. 4. Sanitization: While the wrapper script parses CLI arguments, it does not implement internal sanitization or validation of the manifest's contents before they influence the build process.

lifecyclemodel-automated-builder