The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructions in SKILL.md mandate the execution of node .agents/scripts/precheck.mjs during the initial project validation phase. It also directs the agent to probe the local environment for available binaries and runners (e.g., bun, npm, python3) to determine if generated test commands can be executed.- [DYNAMIC_EXECUTION]: The skill uses templates in references/http-scaffold.md and references/script-templates.md to dynamically generate executable scripts in .agents/docs/acceptance/. These scripts (including .ts, .py, and .spec.ts files) incorporate logic derived from project configuration and user interview responses, which are then presented to the user/agent for execution.- [INDIRECT_PROMPT_INJECTION]: The skill has a large attack surface for indirect injection as it processes untrusted data from the workspace (such as PROJECT.md, package.json, pyproject.toml, and Makefile) and interpolates user-provided strings directly into the authentication headers and test case bodies of the generated scripts.
Ingestion points: Reads project metadata files (PROJECT.md, build configs, test configs) and takes input via the AskUserQuestion tool.
Boundary markers: The skill does not define specific delimiters or instructions to ignore potential commands embedded within the project files it reads.
Capability inventory: The skill has the capability to write files to the filesystem and execute various language runtimes (Node, Python, Bun) and testing frameworks (Playwright, Maestro).
Sanitization: There is no mention of sanitizing or escaping the content retrieved from project files or user interviews before it is placed into executable code blocks.

acceptance