The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses a wide range of Git commands to manage branch state, integration, and documentation movement within the repository.
Evidence: Executes git branch, git status, git checkout, git merge, git rm, git commit, git mv, git reflog, and git config.
[DYNAMIC_EXECUTION]: The skill instructs the agent to execute a shell one-liner to dynamically resolve and verify the plugin's root directory path at runtime.
Evidence: Instructions provided in the 'Rutas de las referencias compartidas' section: PLUGIN_ROOT="${CLAUDE_PLUGIN_ROOT:-}"; [ -n "$PLUGIN_ROOT" ] && PLUGIN_ROOT="$(cd "$PLUGIN_ROOT" 2>/dev/null && pwd)"; echo "PLUGIN_ROOT=$PLUGIN_ROOT".
[INDIRECT_PROMPT_INJECTION]: The skill depends on parsing untrusted external data from markdown files to determine workflow state and task completion.
Ingestion points: Reads task status and metadata from progress.md and other referenced markdown files.
Boundary markers: Uses hidden HTML comments (e.g., ) as state markers, which provides structure but lacks explicit delimiters or instructions to ignore potential malicious content in the same file.
Capability inventory: High-impact capabilities including merging branches, committing changes, moving/deleting files via Git, and executing shell commands.
Sanitization: No sanitization or validation of the markdown content is performed before parsing the hidden state markers.

work-integrate