The Agent Skills Directory

[SAFE]: The skill is composed of technical documentation and reference checklists. It includes high-quality templates for Architecture Decision Records (ADRs), README files, and changelogs designed to improve engineering communication and codebase maintainability.
[INDIRECT_PROMPT_INJECTION]: The skill instructs agents on how to ingest project context and record architectural decisions.
Ingestion points: Agents are directed to analyze repository context, including existing ADRs, project instructions, and configuration files like .adr-dir as seen in SKILL.md.
Boundary markers: The skill relies on standard Markdown structure for documentation templates; it does not explicitly define adversarial boundary markers for data ingestion.
Capability inventory: The primary capabilities involve reading codebase state and writing documentation files.
Sanitization: The security checklist provides specific guidance for developers to implement input validation, parameterized queries, and output encoding in their own applications to prevent injection attacks.
[COMMAND_EXECUTION]: The reference files (references/accessibility-checklist.md, references/performance-checklist.md, references/security-checklist.md) provide examples of standard shell commands for developer tooling, such as npx axe-core, npx lighthouse, and npm audit. These are included as educational examples for manual or automated project auditing.
[EXTERNAL_DOWNLOADS]: The documentation references several well-known developer tools and libraries fetched from official registries (e.g., npm). These references are instructional and target established, reputable software ecosystems.

documentation-and-adrs