The Agent Skills Directory

[COMMAND_EXECUTION]: The skill makes extensive use of the Bash tool to execute loom CLI commands for project orchestration. This includes spawning background daemons (loom run), managing git worktrees (loom worktree), and installing git hooks via loom repair. These operations are consistent with the skill's stated purpose of meta-orchestration.
[PROMPT_INJECTION]: The skill presents an attack surface for indirect prompt injection (Category 8) because it ingests and processes external plan files (PLAN-*.md) to generate task assignments for other sessions.
Ingestion points: Data enters the context via the loom init command and by reading plan markdown files.
Boundary markers: There are no specific instructions in the skill regarding the use of delimiters or 'ignore' instructions for external content.
Capability inventory: The skill has access to powerful tools including Bash, Write, and Edit, which could be targeted by malicious instructions within a plan file.
Sanitization: The instructions mention that the tool performs internal validation for path traversal and metadata structure, which acts as a mitigating factor.

loom-usage