Update to DXD Skills

{
  "name": "agent-native-audit",
  "description": "Audit a codebase for how easily coding agents can understand, change, and verify it. Use for agent-readiness reviews, AI-friendly codebase assessments, or plans to improve agent reliability in a repository.",
  "included_files": [],
  "skill_md_contents": "---\nname: agent-native-audit\ndescription: >-\n  Audit a codebase for how easily coding agents can understand, change, and\n  verify it. Use for agent-readiness reviews, AI-friendly codebase assessments,\n  or plans to improve agent reliability in a repository.\n---\n\n# Agent-Native Codebase Audit\n\n## Goal\n\nProduce an evidence-backed assessment of the conditions that let agents make\nsafe, well-targeted changes, then recommend the highest-leverage improvements.\n\n## When to Use\n\nUse when the user asks to assess agent readiness, score an AI-friendly\ncodebase, or plan changes that make a repository easier for coding agents to\nwork in.\n\n## Evaluation Criteria\n\nAssess the dimensions that matter for the repository at hand:\n\n| Dimension | Look for |\n| --- | --- |\n| Contracts | Types, schemas, explicit boundaries, and helpful failures that make invalid changes hard to write. |\n| Navigation | Predictable structure, naming, ownership, and discoverable canonical paths. |\n| Verification | Fast, meaningful checks that cover relevant behavior and work locally and in CI. |\n| Feedback | Clear commands, actionable errors, and short iteration loops. |\n| Context | Documentation, examples, conventions, and comments that explain intent where code cannot. |\n\nUse scores only when they help the user compare baselines or measure progress.\nIf scoring, state the scale, weighting, uncertainty, and evidence behind each\nrating. Letter grades and fixed score bands are optional presentation choices,\nnot a claim of universal precision.\n\n## Workflow\n\n1. Inspect repository instructions, package and build configuration, CI,\n   verification commands, documentation, and representative source and test\n   paths. Establish the primary languages and the work an agent would perform.\n2. Gather evidence for each relevant criterion. Prefer concrete paths, command\n   output, and examples over inferred maturity.\n3. Run safe local checks when they can confirm a meaningful claim. Distinguish\n   pre-existing failures from the conditions being assessed.\n4. Report strengths, friction, and the evidence for each assessment. Identify\n   the smallest changes likely to improve multiple criteria.\n5. Provide an ordered, executable improvement plan when requested; otherwise\n   offer it as the next step.\n\n## Guardrails\n\n- Never fabricate metrics, coverage, test results, or scores.\n- Never modify source or repair failures during an audit unless the user expands\n  the task.\n- Never judge a repository from a single file or tool configuration.\n- Adapt the assessment to the repository's languages, architecture, and agent\n  environments rather than assuming one IDE or vendor.\n\n## Completion Checklist\n\n- [ ] Findings cite concrete evidence from the repository or checks.\n- [ ] Relevant agent-work criteria were assessed with uncertainty stated.\n- [ ] Recommendations are prioritized by expected leverage and are actionable.\n- [ ] Any score or grade is explained as a chosen reporting convention.\n"
}