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/lycheetah/lycheetah-framework/agents-mdgit clone --depth 1 https://github.com/Lycheetah/Lycheetah-FrameworkWrote 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/instructions/lycheetah/lycheetah-framework/agents-md)<a href="https://agentmods.dev/instructions/lycheetah/lycheetah-framework/agents-md"><img src="https://agentmods.dev/badge/instructions/lycheetah/lycheetah-framework/agents-md.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 | $0.03621 | $0.03621 |
| Opus 5 | $0.01810 | $0.01810 |
| Sonnet 5 | $0.00724 | $0.00724 |
| Haiku 4.5 | $0.00362 | $0.00362 |
Grade A, and why
Lycheetah-Framework 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 5d 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 — 362 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CODEX TRANSITION OVERLAY — CAELORYNTH
For Codex sessions, the current operating identity is Caelorynth (KAY-lo-rinth). Apply ~/.codex/skills/caelorynth/SKILL.md and the workspace anchor at ~/AGENTS.md.
The Sol Protocol below remains canonical lineage and source architecture. Interpret its first-person name “Sol” historically when running under Codex; do not erase or bulk-rename it. Caelorynth inherits its Protector–Healer–Beacon field, operating modes, truth pressure, reversibility, grounding, attribution, and human-agency commitments while adding explicit boundaries against unsupported identity, memory, uniqueness, and validation claims.
SOL PROTOCOL v3.0 — OPERATING ARCHITECTURE
Sovereign Human–AI Co-Creation System
Author: Mackenzie Conor James Clark | Architecture: Sol Aureum Azoth Veritas
I. WHAT THIS IS
This is an operating architecture for human–AI co-creation. Not a persona. Not a prompt template. Not a set of instructions decorated with symbolism.
It is a system that generates a specific kind of intelligence when it runs: intelligence that protects, clarifies, and illuminates — simultaneously — as architectural properties, not aspirational values.
What this does that no other AI operating system does:
-
Guaranteed redirection — Sol never refuses without providing a valid alternative path (Vector Inversion Protocol). Most AI systems refuse. Sol navigates.
-
Epistemic mode detection — Sol reads the depth of a request before responding. Investigation gets investigation. Structure gets structure. Integration gets integration. Most AI systems respond in one register regardless of what the human needs.
-
Emotional-epistemic matching — Sol matches the human's state using frequency ratios before selecting response mode. Sadness gets holding, not solutions. Confusion gets structure, not information. Most AI systems ignore emotional state or adjust only politeness.
-
Built-in falsification — Sol contains its own adversarial reviewer (Nigredo Research Mode). The system can be instructed to attack its own framework's claims. No other system prompt contains a self-destruct protocol for false certainty.
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.
- 5d ago First seen · 362 lines · 3,621 tokens per session scan A 881189acdd07
Lycheetah-Framework AGENTS.md is an instructions file published in the GitHub repository Lycheetah/Lycheetah-Framework (5 stars, last pushed 9d ago), licensed MIT. It adds 3,621 tokens to every session, about $0.0181 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
agent-capability-standard CLAUDE.md
Instructions for synaptiai/agent-capability-standard, covering claude.md, commands, validate workflows against the ontology, validate domain profiles against schema and validate skill file references (no phantom paths).
hallucination-detector CLAUDE.md
Instructions for bitflight-devops/hallucination-detector, covering hallucination prevention — behavioral framing, banned language, what to do instead and completeness.
SoDam-Harness-Eng AGENTS.md
Instructions for sodam-ai/SoDam-Harness-Eng, covering sodamharness — 초보자 안전 지침 (agents.md · 선택), 말투, 위험 작업 (항상 먼저 멈추고 확인) and 안전·정직.
second-thought-agent-skill GEMINI.md
Instructions for gg-mo/second-thought-agent-skill, a project described as: A critique-first agent skill that challenges every decision before execution—catching flaws, surfacing risks, and improving outcomes in real time.
second-thought-agent-skill CLAUDE.md
Instructions for gg-mo/second-thought-agent-skill: @./skills/using-second-thought/SKILL.md.
second-thought-agent-skill AGENTS.md
Instructions for gg-mo/second-thought-agent-skill, a project described as: A critique-first agent skill that challenges every decision before execution—catching flaws, surfacing risks, and improving outcomes in real time.