The Agent Skills Directory

[COMMAND_EXECUTION]: The skill makes extensive use of the Bash tool to perform project setup and maintenance tasks.
Evidence:
Executes environment and code quality checks using tools like uv, ruff, mypy, and pytest.
Interacts with the GitHub CLI (gh) to check authentication status, refresh tokens, and manage project boards.
Runs various hygiene tools including gitleaks for secret scanning, shellcheck for script linting, and actionlint for GitHub Actions validation.
Orchestrates paper-specific workflows using tex-fmt and custom check scripts.
[EXTERNAL_DOWNLOADS]: The skill downloads an external template repository to initialize the slides component.
Evidence:
Executes git clone https://github.com/a-green-hand-jack/progress-slides.git to fetch a slide deck template from the author's own repository.
[INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection as it processes user-provided summaries and interacts with existing code/paper repositories.
Ingestion points: User-provided research summaries (Step 1) and content from existing repositories when connected (Step 5) are ingested into the agent's context.
Boundary markers: The skill uses structured YAML files (memory/project.yaml) and dedicated markdown guidance files (AGENTS.md, CLAUDE.md) to define operational boundaries and project-specific policies.
Capability inventory: The agent has access to Read, Write, Edit, Bash, and Glob tools across the project root and component repositories.
Sanitization: No explicit automated sanitization of external data is mentioned; security relies on the agent adhering to the instructions provided in the guidance files and human review of generated diffs.

project-init