The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it directs the agent to ingest and act upon data from project-level files and task notes.
Ingestion points: The agent reads the Environment Status Cache from project.md (Step 0) and task context from bd show (Step 4).
Boundary markers: No explicit delimiters or instructions are provided to the agent to treat this external data as untrusted or to ignore embedded instructions.
Capability inventory: The agent has the capability to execute shell commands, install software, and modify project files.
Sanitization: There is no evidence of sanitization or schema validation performed on the contents of the project.md file before it is processed by the agent.
[EXTERNAL_DOWNLOADS]: The skill automates the installation of the BeadBoard coordination system by cloning its repository from GitHub and installing the beads-cli package from the NPM registry. These resources are authored by the skill's vendor.
[COMMAND_EXECUTION]: Several utility scripts use spawnSync and execFile to invoke CLI tools for environment setup and coordination. Notably, it configures a mail delegate in the bd tool which executes a local shim script to handle agent communications.
[REMOTE_CODE_EXECUTION]: The skill facilitates the installation and execution of external code through global NPM package installation and repository cloning as part of its initialization and execution runbook.

beadboard-driver