Borrowing it
Nothing to install: this file belongs to thalesholleben/ccx. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/thalesholleben/ccx/master/AGENTS.mdgit clone --depth 1 https://github.com/thalesholleben/ccxWrote 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/thalesholleben/ccx/agents-md)<a href="https://agentmods.dev/instructions/thalesholleben/ccx/agents-md"><img src="https://agentmods.dev/badge/instructions/thalesholleben/ccx/agents-md/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/instructions/thalesholleben/ccx/agents-md"><img src="https://agentmods.dev/badge/instructions/thalesholleben/ccx/agents-md.svg" alt="Reviewed on agentmods" width="80" 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 5.1 | $0.00783 | $0.00783 |
| Opus 5.5 | $0.00313 | $0.00313 |
| Sonnet 5.5 | $0.00157 | $0.00157 |
| Haiku 4.5 | $0.00078 | $0.00078 |
Grade A, and why
ccx 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 1mo 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 — 59 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CCX contributor instructions
CCX is a local account fleet for official Claude Code and Codex CLIs. Read
README.md and docs/features/fleet.md first. Before invoking another agent,
read skills/ccx/SKILL.md; this does not authorize extra delegation.
Map
ccx-fleet.py,fleet/cli.py: primary CLI.ccx.py fleetis an alias.fleet/store.py: SQLite state, validation, deduplication and schema migration.fleet/scheduler.py: quota, reservations, account selection, workspace exclusion.fleet/providers.py: official CLI adapters, isolated login, quota parsing/refresh.fleet/service.py,runner.py,processes.py: admission, reconciliation and process ownership.fleet/panel.py,presentation.py,tray.py: Tk dashboard and Windows tray.skills/ccx: generic distributable skill;scripts/install-skill.py: dual installation.ccx.py,ccx_codex.py,ccx_runtime.py: compatibility modules and reused primitives. Global monitor/bridge entry points are retired; installers accept uninstall only. ccx.py stats and ccx_codex.py status route to the fleet.
Validate
python scripts/check.py runs offline regression suites and the installer test.
Desktop changes also require python tests/panel-smoke.py and
python tests/tray-smoke.py on interactive Windows. No real accounts in tests.
Pillow is optional for screenshots/icon exports. Runtime must stay stdlib-only.
Invariants
- One account pool per identity; a separately authenticated OAuth home per worker. Do not clone grants, alter global client login, or open managed profiles externally.
- Login, refresh and execution hold the same OS worker lock. Credential writes preserve unrelated fields and replace atomically; corrupt JSON never means empty.
- Admission and reservation share one transaction. No numeric global/per-cell cap; schema 3 keeps retired max_active columns at zero for compatibility only.
- Unknown/stale quota or expired reset requires measurement, never projected zero. HTTP 429 is a measurement error; 401/403 block that profile for login.
- Session and weekly weights differ. Keep model-scoped windows and conservative reservations. Completed reservations need a delayed fresh measurement to release.
- One ticket claim, one runner. Ambiguous failures do not replay prompts. Stop the service without killing live runners; close the panel without stopping service.
- Overlapping paths exclude any job when one is a writer. Keep ownership until the process tree exits, even after a terminal event.
- Preserve permissions, model, effort, recursion guards and explicit account affinity.
- Errors/logs must not reveal credentials or private prompts. Never commit real state.
- Schema 4 adds display_name without changing IDs or credentials. UI/CLI registration generates an internal ID; label edits preserve jobs, reservations and profile paths.
- Schema 5 separates provider plan weights. Old Codex pro x1 becomes plus; calibrated old pro weights become custom unchanged. Never assume Claude Pro equals Codex Pro.
- Account removal requires idle unlocked profiles and no queued affinity. Preserve task history/results; delete only validated local profile paths. Report cleanup failures.
- Protocol transports preserve schema, permissions, tools and recursion guards. Cancellation on timeout must confirm runner exit before another dispatch.
- Registration UI has name/provider/plan only. The dashboard monitors accounts and jobs; it does not create tasks.
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.
- 1mo ago Changed · -84 lines · -1,538 tokens per session 05eb91b9b61e
- 1mo ago Changed · +41 lines · +763 tokens per session 7b77f9ccc726
- 1mo ago First seen · 102 lines · 1,558 tokens per session scan A 0885909629c9
ccx AGENTS.md is an instructions file published in the GitHub repository thalesholleben/ccx (9 stars, last pushed 1mo ago), licensed MIT. It adds 783 tokens to every session, about $0.0031 per session on Opus 5.5. 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
vscode buildNext.instructions.md
Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).
next.js AGENTS.md
AGENTS.md instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.
vscode oss-third-party-notices.instructions.md
Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, cache eligibility and rollout and applying the notice (cutover).
langchain AGENTS.md
AGENTS.md instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.
vscode copilot-instructions.md
Copilot instructions for microsoft/vscode, covering vs code copilot instructions, project overview, root folders, core architecture (src/ folder) and built-in extensions (extensions/ folder).
deepseek-harness AGENTS.md
AGENTS.md instructions for deepseek-ai/deepseek-harness, covering agents.md, pre-stable apis and released session data, repository layout, commands and host sandbox failures.