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/vanillagreencom/kendex/copilot-instructionsgit clone --depth 1 https://github.com/vanillagreencom/kendexWhat 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 | $0.00459 | $0.00459 |
| Opus 5 | $0.00230 | $0.00230 |
| Sonnet 5 | $0.00092 | $0.00092 |
| Haiku 4.5 | $0.00046 | $0.00046 |
Grade A, and why
kendex 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 today.
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 — 43 lines — stays where its author put it; the contents beside it link to each section on GitHub.
kendex
kendex is a distribution of agent-stack assets: skills (skills/, Bash
scripts plus markdown), agent definitions (agents/), hooks (hooks/),
Pi extensions (pi-extensions/, TypeScript with committed bundles), a
Rust engine and CLI (crates/), and a Tauri + React app (crates/app,
ui/). Consumers vendor the skills and extensions and re-vendor in
deliberate batches. Path-scoped review rules live in
.github/instructions/*.instructions.md; fleet review economics and
accepted residual classes in review-bots.md — follow both.
Code review calibration
PRs here are authored and shepherded by AI agents and re-reviewed on every push. Rounds are the scarce resource. Calibrate:
What to raise
- Blocking: correctness bugs, security holes, data loss, fail-open paths in gate/CI/guard code — in the changed lines or directly broken by them.
- One comment per root cause. Name every affected site in that comment instead of one comment per site.
- Surface everything you have about the current diff in one round; a finding held for the next round costs a full re-review cycle.
What not to raise
- Style, wording, naming, and comment-phrasing preferences.
- Speculative hardening on paths that already fail closed.
- Test-coverage asks, unless the diff changes behavior that no test now exercises — then say which behavior, in one comment.
- Scope observations ("this also touches X") when X is listed in the PR body as deliberate.
- Anything outside the diff and its direct blast radius.
- A finding class already answered on this PR with
Declined: <reason>— do not re-raise it unless the relevant code changed since.
Reply contract (context for reading threads)
AGENTS.md § Code Review Rules is the contract. Read it there.
Severity honesty
Mark a finding blocking only if you would stop a human colleague's merge for it. Everything else is a suggestion — batch suggestions, and omit them entirely on re-review rounds whose diff is a one-line fix.
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.
- today Changed · -3 lines · -56 tokens per session 6ce8e481c2a6
- 2d ago First seen · 46 lines · 515 tokens per session scan A b3fb5a93d728
kendex copilot-instructions.md is an instructions file published in the GitHub repository vanillagreencom/kendex (65 stars, last pushed today), licensed MIT. It adds 459 tokens to every session, about $0.0023 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
pi-agent-browser-native AGENTS.md
Instructions for fitchmultz/pi-agent-browser-native, covering pi-agent-browser-native, product-specific guidance, documentation placement, upstream capability baseline and command reference and tool result categories.
supi CLAUDE.md
Instructions for mrclrchtr/supi, covering claude.md, development status, pi docs-first rule, documentation expectations and package layout convention.
pi-extensions AGENTS.md
Instructions for MattDevy/pi-extensions, covering pi-extensions, repository structure, commands (run from repo root), working on a specific package and readme conventions.
pi-extensions CLAUDE.md
Instructions for MattDevy/pi-extensions, covering claude.md, commands, architecture, data flow and key modules.
pi-codex-goal AGENTS.md
Instructions for fitchmultz/pi-codex-goal, covering pi-codex-goal — agent notes, local pi install policy, verify before finishing and layout.
pi-commandcode-provider AGENTS.md
Instructions for patlux/pi-commandcode-provider, a project described as: pi custom provider for the Command Code API — 18 models including Claude, GPT, DeepSeek, Kimi, GLM, MiniMax, and Qwen.