The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from repository files such as CLAUDE.md, AGENTS.md, and docs/agents/triage-labels.md during its initial exploration phase.
Ingestion points: SKILL.md reads root-level configuration files and docs/agents/ documentation to determine the current repository state.
Boundary markers: There are no specific delimiters or instructions to ignore embedded commands when reading these files.
Capability inventory: The skill uses shell commands via git, gh (GitHub CLI), and glab (GitLab CLI), and it has write access to the filesystem.
Sanitization: The instructions mandate presenting findings to the user and obtaining confirmation before performing write operations or executing non-read commands, which serves as a human-in-the-loop defense mechanism.
[COMMAND_EXECUTION]: The skill provides instructions for the agent to execute shell commands for project management.
Evidence: issue-tracker-github.md and issue-tracker-gitlab.md include bash scripts and command patterns for managing issues, labels, and pull requests using the gh and glab CLI tools.
Context: These operations are standard for the skill's purpose of automating the setup of a repository's engineering workflow.

bootstrap