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.
npx agentmods add instructions/microsoft/promptkit/copilot-instructionsgit clone --depth 1 https://github.com/microsoft/PromptKitWhat 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 | $0.01144 | $0.01144 |
| Opus 5 | $0.00572 | $0.00572 |
| Sonnet 5 | $0.00229 | $0.00229 |
| Haiku 4.5 | $0.00114 | $0.00114 |
Grade A, and why
PromptKit copilot-instructions.md 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 yesterday.
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 — 139 lines — stays where its author put it; the contents beside it link to each section on GitHub.
PromptKit — Copilot Instructions
How to Use This Repository
When a user asks you to perform a task that this library supports (writing
requirements, investigating bugs, reviewing code, etc.), use the
/promptkit skill (or read and execute bootstrap.md). It is the entry
point that guides component selection, parameter gathering, and prompt
assembly. Do not try to answer the task directly — assemble the right
prompt first.
What This Repository Is
PromptKit is a composable prompt library, not a software application. There is no build system, no runtime, and no application code. Every file is Markdown or YAML — the "code" is prose that LLMs consume.
The core loop: bootstrap.md reads manifest.yaml to discover components,
then assembles them into task-specific prompts from five composable layers:
personas → protocols → formats → taxonomies → templates.
Architecture
Templates are the orchestration layer. Each template declares which persona,
protocols, and format to compose via YAML frontmatter. The bootstrap engine
reads these declarations from manifest.yaml and assembles a single coherent
prompt in this order:
- Identity (persona)
- Reasoning Protocols (one or more, in order)
- Classification Taxonomy (if applicable)
- Output Format (structure rules)
- Task (template with
{{param}}placeholders filled) - Non-Goals
Templates can chain via pipelines — each declares input_contract /
output_contract with artifact types (e.g., requirements-document →
design-document). The document-lifecycle pipeline chains
requirements → design, plus validation consuming requirements and
optionally design.
Component Conventions
All components use kebab-case filenames matching their name field.
Protocols have three categories with different scoping rules:
guardrails/— Cross-cutting, apply to all tasks (anti-hallucination, self-verification, operational-constraints)analysis/— Domain/language-specific checks (memory-safety-c, security-vulnerability). Language-specific protocols are separate files, not conditional blocks.reasoning/— Systematic reasoning approaches (root-cause-analysis, requirements-elicitation)
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.
- yesterday First seen · 139 lines · 1,144 tokens per session scan A 41b837845ec6
PromptKit copilot-instructions.md is an instructions file published in the GitHub repository microsoft/PromptKit (102 stars, last pushed 12d ago), licensed MIT. It adds 1,144 tokens to every session, about $0.0057 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 instructions, from other repositories
claude-code-settings copilot-instructions.md
Instructions for feiskyer/claude-code-settings, covering claude.md, environment setup, required dependencies, configuration and skills.
docs-mcp-server AGENTS.md
Instructions for arabold/docs-mcp-server, covering agent instructions for docs-mcp-server, repository context, development workflow, key commands and git workflow.
AI-Agents-Orchestrator AGENTS.md
Instructions for hoangsonww/AI-Agents-Orchestrator, covering agents.md — instructions for ai coding agents, project overview, build & test, code style and architecture boundaries.
agent-rules testing.instructions.md
Instructions for lirantal/agent-rules, covering testing guidelines for node.js applications, general testing principles, guiding principles, descriptive test naming (three-part structure) and structure tests by the aaa pattern.
apex-accelerator agent-skills.instructions.md
Guidelines for creating high-quality Agent Skills for GitHub Copilot.
apex-accelerator context-optimization.instructions.md
Context window optimization rules for agent definitions, skills, and instruction files.