mirror of
https://github.com/github/awesome-copilot.git
synced 2026-08-07 09:38:45 +00:00
Aaron Powell
a7fdcd5006
* feat: migrate plugins and extensions to Agent Plugins v1.0.0 spec - Add \ to all 69 curated plugin manifests - Migrate all 18 extension manifests: add \, move logo into xtensions.com.github.copilot.logo namespace, remove top-level logo and string xtensions: '.' - Update eng/validate-plugins.mjs: require \, validate namespace-keyed extensions object for canvas extensions, widen name pattern to allow dots (spec §5.5, max 64 chars) - Update eng/materialize-plugins.mjs: emit spec-clean served manifests (only spec fields: \, name, version, description, author, homepage, repository, license, keywords, extensions) - Update eng/generate-website-data.mjs: read logo from namespace with fallback to top-level logo for compatibility - Update eng/create-plugin.mjs: scaffold emits \ - Add .github/workflows/validate-plugins.yml: blocking CI for PRs touching plugins/** or extensions/** - Add spec compliance check to external plugin quality gates: non-blocking warnings with ✅/⚠️/🛑 emoji legend - Update AGENTS.md: document new extension manifest shape, add \ to plugin checklist Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com> Copilot-Session: 8f3a88cb-e01e-4760-8125-460490dc1a76 * refactor: consolidate canvas extension plugins - Move all extension plugin manifests from extensions/<name> to plugins/<name> - Keep extensions/<name> as reusable source only - Remove standalone extension discovery from marketplace and website plugin catalogs - Auto-bundle same-name extension sources during materialization - Add build-only extensions.json references for sharing extensions across plugins - Remove x-awesome-copilot extension metadata support - Update validation and contributor documentation Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com> Copilot-Session: 8f3a88cb-e01e-4760-8125-460490dc1a76 * feat: add canvas extension scaffolding skill - Add repo-local skill for creating canvas extension sources - Generate spec-compliant plugin manifests under plugins/ - Support registering reusable extensions with multiple plugins - Remove guidance for extension-local plugin manifests and custom fields Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com> Copilot-Session: 8f3a88cb-e01e-4760-8125-460490dc1a76 * fix: align extension namespaces with current guidance - Use each extension ID as its manifest namespace key - Update validation and website generation to resolve extension-specific namespaces - Upsert plugin validation PR comments using the existing repository pattern Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com> Copilot-Session: 8f3a88cb-e01e-4760-8125-460490dc1a76 * fix: use Copilot extension namespace - Adopt com.github.copilot for all canvas extension manifests - Require the namespace during validation and website generation - Update extension scaffolding guidance Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com> Copilot-Session: 8f3a88cb-e01e-4760-8125-460490dc1a76 * docs: regenerate plugin catalog after merge Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com> Copilot-Session: 8f3a88cb-e01e-4760-8125-460490dc1a76 * refactor(plugins): move manifests to plugin roots Use root plugin.json manifests and namespaced extension directories throughout local tooling, validation, generation, and contributor documentation. Restore materialize-plugins.mjs line breaks so the source remains readable in GitHub. Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com> Copilot-Session: 8f3a88cb-e01e-4760-8125-460490dc1a76 * feat(plugins): migrate manifests to namespaced composition Move repository composition metadata under com.github.awesome-copilot, materialize reusable extensions into the plugin extensions directory, and improve contributor and PR validation guidance. Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com> Copilot-Session: 8f3a88cb-e01e-4760-8125-460490dc1a76 * fix(validation): address plugin review findings Restore executable build scripts, validate namespaced manifests and hook directories, improve README item counts, and manage validation comments across reruns. Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com> Copilot-Session: 8f3a88cb-e01e-4760-8125-460490dc1a76 * Potential fix for pull request finding Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com> --------- Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com> Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com> Copilot-Session: 8f3a88cb-e01e-4760-8125-460490dc1a76
acreadiness-cockpit
Drive Microsoft AgentRC from Copilot chat. Frames every interaction inside AgentRC's Measure → Generate → Maintain loop.
What's in the plugin
Custom agent
| Agent | What it does |
|---|---|
@ai-readiness-reporter |
Runs agentrc readiness --json, interprets every result against the 9-pillar / 5-level model, then renders a self-contained reports/index.html from a fixed HTML/CSS template so every user gets an identically styled dashboard. Honours policies (disabled criteria, overrides, pass-rate thresholds) and surfaces extras separately. |
Skills
| Skill | Step | What it does |
|---|---|---|
/acreadiness-assess |
Measure | Runs the readiness scan and hands off to @ai-readiness-reporter to produce the static HTML dashboard. Accepts --policy <path-or-pkg> and --per-area. |
/acreadiness-generate-instructions |
Generate | Wraps agentrc instructions. Default output is .github/copilot-instructions.md (Copilot-native). Asks flat vs nested. For monorepos, also emits per-area .github/instructions/<area>.instructions.md files with applyTo globs. |
/acreadiness-policy |
Maintain | Pick, scaffold, or apply an AgentRC policy. Knows the schema (criteria.disable, criteria.override, extras, thresholds), the impact-weight table, and CI gating with --fail-level. |
What gets produced
reports/index.html — a single self-contained HTML file rendered from a fixed template (skills/acreadiness-assess/report-template.html) so every user gets an identical look & feel. It contains:
- Maturity badge (L1–L5) and overall score / grade (A–F)
- Pass-rate vs threshold (when a policy sets one)
- Maturity progression table
- Active policy summary (disabled/overridden criteria, threshold)
- Repo Health breakdown (8 pillars), each with an AI relevance badge (High/Medium/Low), what it measures, why it matters for AI, current state, recommendation
- AI Setup breakdown (AI Tooling pillar)
- Extras (informational only — agents-doc, pr-template, pre-commit, architecture-doc)
- Prioritised Remediation Plan (🔴 Fix First / 🟡 Fix Next / 🔵 Plan)
- Embedded raw AgentRC JSON for reuse
Prerequisites
- Node.js 20+ on PATH (required by AgentRC)
- VS Code with Copilot agent plugins enabled
Usage
In Copilot chat:
/acreadiness-assess # measure → reports/index.html
/acreadiness-assess --policy ./policies/strict.json
/acreadiness-generate-instructions # asks flat or nested
/acreadiness-generate-instructions --strategy flat
/acreadiness-generate-instructions --strategy nested
/acreadiness-generate-instructions --areas # per-area applyTo files
/acreadiness-policy new my-policy
@ai-readiness-reporter
Flat vs nested instructions
| Flat (default) | Nested | |
|---|---|---|
| Hub file | .github/copilot-instructions.md |
.github/copilot-instructions.md |
| Detail files | — | .github/instructions/<topic>.instructions.md (each with applyTo glob) |
| Best for | Small / medium repos, single stack | Large or multi-stack repos, monorepos |
| Token cost | Whole file always loads | VS Code only loads topics whose applyTo matches |
When the main output is .github/copilot-instructions.md, the skill rewrites AgentRC's nested output to VS Code's native .instructions.md layout (which Copilot auto-discovers). With --output AGENTS.md, nested keeps AgentRC's default .agents/ layout for agent-agnostic tooling.
Concepts (cheat sheet)
- Maturity: L1 Functional → L2 Documented → L3 Standardized → L4 Optimized → L5 Autonomous
- Pillars (Repo Health): Style · Build · Testing · Docs · Dev Environment · Code Quality · Observability · Security
- Pillars (AI Setup): AI Tooling
- Impact weights: critical 5 · high 4 · medium 3 · low 2 · info 0
- Grades: A ≥ 0.9 · B ≥ 0.8 · C ≥ 0.7 · D ≥ 0.6 · F < 0.6
License
MIT