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/cognesy/instructor-php/agents-mdgit clone --depth 1 https://github.com/cognesy/instructor-phpWhat 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 | $0.03583 | $0.03583 |
| Opus 5 | $0.01792 | $0.01792 |
| Sonnet 5 | $0.00717 | $0.00717 |
| Haiku 4.5 | $0.00358 | $0.00358 |
Grade A, and why
instructor-php AGENTS.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 — 281 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Key Reference Files
Monorepo Root
- CONTRIBUTING.md - Development workflows, package management, and contribution guidelines
- CONTENTS.md - Project structure and package overview
- README.md - Main project documentation
- HARD REFERENCE -
CONTRIBUTING.mdis the source of truth for QA execution (composer qa,qa:*) and contribution workflow details.
Individual Subpackages
Each package in packages/ may contain:
- README.md - Package-specific documentation
- OVERVIEW.md - Package overview and architecture
- INTERNALS.md - Implementation details and internal structure
- CHEATSHEET.md - Quick reference for package usage
Code Style
- Use strict types and type hints for arguments and return values everywhere
- Do not nest control structures (if, loops) beyond 1 level, exceptionally 2 levels
- Use
matchfor complex conditionals - avoidif/else ifchains orswitchstatements; avoid ternary operators - Prefer immutable data structures and functional programming paradigms
- Avoid using arrays as collections - use dedicated collection classes
- Avoid exceptions for control flow - do not wrap everything in try/catch; either let exceptions bubble up or use monadic Cognesy\Utils\Result\Result for error handling if error needs to be handled on the same level
- Use namespaces and PSR-4 autoloading
Agents Package Style
- Data classes are
readonly final class— all properties readonly, all mutations return new instances viawith*()methods - Interfaces use
Can-prefix (CanUseTools,CanEmitAgentEvents,CanInterceptAgentLifecycle) - Hooks use
Hooksuffix (StepsLimitHook,ErrorPolicyHook) - Accessors have no prefix:
state(),execution(),currentStep()— notgetState() - Mutators use
withprefix and returnself:withState(),withCurrentStep() - Named constructors for defaults:
::empty(),::fresh(); for hydration:::fromArray() - Minimal docblocks — only on class/interface declarations; methods are self-documenting via types and naming
- Method order: constructors → lifecycle/transitions → accessors → mutators → private helpers → serialization (
toArray/fromArray) - Enums are string-backed for serialization; put logic (priority, comparison) as methods on the enum itself
- Use
match(true) { ... }for multi-condition branching - Use
$this->field ?? Default::empty()pattern ("ensure pattern") to provide defaults instead of null-checking
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 · 281 lines · 3,583 tokens per session scan A d597dafda963
instructor-php AGENTS.md is an instructions file published in the GitHub repository cognesy/instructor-php (326 stars, last pushed 2d ago), licensed MIT. It adds 3,583 tokens to every session, about $0.0179 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
optillm CLAUDE.md
Instructions for algorithmicsuperintelligence/optillm, covering claude.md, project overview, core architecture, main components and development commands.
AI-Gateway AGENTS.md
Instructions for Azure-Samples/AI-Gateway, covering agents.md, directory structure, labs/, modules/ and shared/.
iii AGENTS.md
AGENTS.md instructions for iii-hq/iii, covering agents.md, commands, setup, build and test.
remind AGENTS.md
Instructions for sandst1/remind, covering remind - development guide for ai agents, project overview, architecture, key abstractions and data models (models.py).
sdk-typescript AGENTS.md
Instructions for strands-agents/sdk-typescript, covering agent development guide - strands typescript sdk, purpose and scope, directory structure, directory purposes and development workflow for agents.
skills AGENTS.md
Instructions for agentsphp/skills, covering agents.md, project overview, key features, commands and requirements.