The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill relies on the @web42/stask package, which is fetched from the NPM registry during execution.
[REMOTE_CODE_EXECUTION]: Uses npx to execute the @web42/stask tool, resulting in the execution of remotely downloaded code.
[COMMAND_EXECUTION]: Executes shell commands such as npx for task lifecycle operations and git for managing worktrees and branches.
[DATA_EXFILTRATION]: Transmits task metadata, progress updates, and results to Slack using the chat.postMessage API. While Slack is a well-known service, this involves sending project-specific data externally.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it processes task specifications (specs/*.md) and Slack thread content which could contain malicious instructions.
Ingestion points: Reads external task specifications from files and monitors Slack thread replies for updates.
Boundary markers: Absent; the skill does not define specific delimiters or provide instructions to the agent to disregard instructions embedded within the task specs.
Capability inventory: The agent can execute shell commands (npx, git), interact with the filesystem (via worktrees), and communicate with the Slack API.
Sanitization: No validation or sanitization of the specification content is documented.

stask-general