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/livetrack-x/spec-driven-ai-development/copilot-instructionsgit clone --depth 1 https://github.com/LiveTrack-X/spec-driven-ai-developmentWhat 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 | $0.01215 | $0.01215 |
| Opus 5 | $0.00607 | $0.00607 |
| Sonnet 5 | $0.00243 | $0.00243 |
| Haiku 4.5 | $0.00121 | $0.00121 |
Grade A, and why
spec-driven-ai-development 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 2d 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 — 113 lines — stays where its author put it; the contents beside it link to each section on GitHub.
SDAD Protocol
Status: Active Scope: GitHub Copilot project instructions
SDAD (SPEC-Directed AI Development) is a method-agnostic, repository-local operating protocol for AI-assisted development, not a coding method, agent runtime, or substitute for owner evidence.
Fast Start
Use progressive disclosure:
- Read
sdad-state.yamlfor scale, packet, gates, validation, and eligible routes. - Read
docs/INDEX.mdas a router, not a reading list. - Inspect current source and tests. Current intent selects only the routed path, heading, active section, or targeted match; membership does not mean read the whole file.
- Load history/optional evidence on demand, including relevant
docs/Repository-Operating-Rules.mdheadings for policy, risk, or release.
Controls And Route
Use Plan -> Route -> Implement -> Verify -> Report. Owner gates and handoffs are triggered branches. Mark an inapplicable read-only/planning phase N/A, explain why, and never claim evidence from it.
Controls are scale, execution_scope: unit | packet, and owner gates: Mini uses
one unit; Standard one packet; Full adds gates. Ask-first means no authorization.
Multi-packet needs an approved list, never session scope. routed_docs permits
selection, not authority or a full read.
Infer review/audit, implement/fix, docs, handoff/resume, or protected-action intent from the request and repository. Inspect evidence before asking; ask at most one blocking question only when an unresolved fact changes direction, authority/reference role, execution boundary, gate, or claim. Review/audit is read-only unless the owner authorizes changes. Resume only the declared handoff.
When a current owner instruction redirects work, stop affected local and delegated work, re-enter Plan/Route, and reconcile SPEC/state. Old-boundary outputs stay stale until revalidated. A clear imperative authorizes only its named action/boundary; persist it and do not re-ask. A question, hypothetical, quotation, or negation is not authorization. Evidence, prerequisites, tool policy, and other gates remain.
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.
- 2d ago First seen · 113 lines · 1,215 tokens per session scan A ee914a5ebaa5
spec-driven-ai-development copilot-instructions.md is an instructions file published in the GitHub repository LiveTrack-X/spec-driven-ai-development (4 stars, last pushed 1mo ago), licensed MIT. It adds 1,215 tokens to every session, about $0.0061 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 2026-08-31.
Other instructions, from other repositories
blueprint AGENTS.md
Instructions for owainlewis/blueprint, covering blueprint repository policy, principles, writing, phases and workflow for code changes.
blueprint CLAUDE.md
Instructions for owainlewis/blueprint, a project described as: The best agent skills in the world for software development.
specwright AGENTS.md
Instructions for Obsidian-Owl/specwright, covering specwright, workflow, anchor documents, architecture and operator notes.
specwright CLAUDE.md
Instructions for Obsidian-Owl/specwright, a project described as: Craft quality software with AI discipline. Spec-driven development plugin for Claude Code and Opencode — quality gates, adversarial testing, and evidence capture.
specwright aikido_rules.instructions.md
Aikido MCP Security Rules.
aider-desk AGENTS.md
AGENTS.md instructions for hotovo/aider-desk, covering agents.md, common commands, development, type checking and linting and formatting.