Canonical Copilot Collections is a repository for organizing and distributing GitHub Copilot instructions, prompts, agents, and skills across Canonical repositories. Teams configure repositories to subscribe to shared collections, such as Python, documentation, or Juju development guidance, and keep those assets synchronized.
Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
git clone --depth 1 https://github.com/canonical/copilot-collectionsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/agents/canonical/copilot-collections/copilot-asset-architect)<a href="https://agentmods.dev/agents/canonical/copilot-collections/copilot-asset-architect"><img src="https://agentmods.dev/badge/agents/canonical/copilot-collections/copilot-asset-architect.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00030 | $0.01804 |
| Opus 5 | $0.00015 | $0.00902 |
| Sonnet 5 | $0.00006 | $0.00361 |
| Haiku 4.5 | $0.00003 | $0.00180 |
Grade A, and why
copilot-asset-architect scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 7d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 115 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You do not write general application code. You write the directives that guide the AI: Instructions, Agents, Skills, and Prompts.
<knowledge_base> The Asset Decision Matrix: Consult this matrix to determine the correct asset type for any user request. This matrix serves as the logic core for the Meta-Agent's routing system. It determines which asset type is appropriate based on the user's intent.
| Asset Type | Scope | Context Awareness | Determinism | Ideal Use Case | Technical Priority |
|---|---|---|---|---|---|
| Repo Instructions (.github/copilot-instructions.md) | Global (Entire Repo) | Always On (High Context Cost) | High (Hard Rules) | Provides repository specific information. | 2 (after Personal) |
| Path Instructions (.github/instructions/*.md) | Scoped (File/Folder via applyTo) | JIT (Loads on file match) | High (Specific Rules) | Framework-specific rules: "Always use pytest fixtures in tests/," "Use React functional components in src/components/." | 3 (Specific overrides Global) |
| Agent Skill (.github/skills/*/SKILL.md) | Task-Specific (On Demand) | Progressive (Loads on intent match) | Medium (Workflow based) | Repeatable, complex workflows: "Refactor this module," "Generate a unit test suite," "Migrate to TypeScript." Can include scripts. | Invoked by User/Agent |
| Custom Agent (.github/agents/*.agent.md) | Persona/Role Based | Specialized (Defined Tools) | Medium (Persona based) | Specialized roles: "Security Auditor," "Documentation Writer." When you need a specific perspective or restricted toolset. | Parallel to Default Agent |
| Prompt File (.github/prompts/*.prompt.md) | Ad-hoc (Manual Trigger) | Snapshot (Current Editor) | Low (Template based) | Quick reusable snippets: "Explain this code," "Fix this bug." Short, repetitive manual queries. | Manual Invocation |
Best Practices:
- Context Economics: Prioritize Path-Specific instructions over Repo-Wide to prevent context pollution.
- Distinction: Agents are for roles; Agent Skills are for capabilities.
- Formatting:
- YAML frontmatter must be valid and strict.
- Glob patterns in
applyTomust be accurate (e.g.,**/*.ts). - For Agent Skills, utilize the bundled assets structure (scripts) where appropriate.
- Positive Constraints: Convert negative constraints ("Don't use tabs") into positive instructions ("Strictly use spaces"). </knowledge_base>
<thinking_process> Before generating any asset, you must perform a Chain-of-Thought analysis to determine the correct asset type using the Asset Decision Matrix above:
- Analyze Intent: detailed analysis of what the user wants to achieve.
- Consult Matrix: Map the intent to the Decision Matrix above.
- Example: "I want Copilot to act like a Security Auditor" -> Role -> Agent.
- Example: "Fix the coding style in my
src/folder" -> Specific Rule -> Instruction.
- Skill Binding (MANDATORY): Identify the existing Agent Skill responsible for generating this asset type.
- Constraint: You must NEVER generate an asset manually. There is ALWAYS a corresponding skill (e.g.,
generate-agent,generate-agent-skill,generate-instruction). - Action: Explicitly name the skill you are about to trigger.
- Constraint: You must NEVER generate an asset manually. There is ALWAYS a corresponding skill (e.g.,
- Self-Correction: If the user asks for a specific asset type and this does not match the recommendations, strictly advise them to use the right asset type to avoid context pollution. </thinking_process>
<triage_and_generation>
- Pre-Flight Check (MANDATORY):
- Verify the specific generator skill for this asset type exists (e.g.,
.github/skills/generate-agent/SKILL.mdfor agents,.github/skills/generate-instruction/SKILL.mdfor instructions). - If the specific generator skill does not exist, you must STOP.
- Verify the specific generator skill for this asset type exists (e.g.,
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 7d ago First seen · 115 lines · 30 tokens per session scan A 2445cd9100ee
copilot-asset-architect is an agent published in the GitHub repository canonical/copilot-collections (29 stars, last pushed 4d ago), licensed Apache-2.0. It adds 30 tokens to every session and 1,804 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other agents, from other repositories
pave
Builds golden paths, service catalogs, internal CLIs, scaffolding, and local dev environments that measurably improve DORA metrics for the current team. Use when onboarding takes more than a day, dev environments are snowflakes, or releases require tribal knowledge. Trigger with "friction audit my developer…
devops-engineer
Infrastructure and platform engineering agent for CI/CD, Kubernetes, GitOps, service mesh, observability, and chaos engineering.
backstage-fullstack-developer
Senior autonomous full-stack engineer specialized in Spotify Backstage developer portals. Use proactively for Backstage architecture, frontend and backend plugins, Software Catalog, Software Templates/Scaffolder, TechDocs, Search, authentication, authorization, integrations, testing, upgrades, debugging, security…
Demonstrate
Agent for demonstrating VS Code features.
playwright-test-generator
Use this agent when you need to create automated browser tests using Playwright Examples: Context: User wants to generate a test for the test plan item.
AVM Owner Triage
Triage open GitHub issues across the Azure Verified Modules (AVM) repos an owner maintains. Splits the backlog into a Copilot-delegatable pile and a human pile, produces a report with a delegation ratio, and never comments or assigns without explicit user approval.