Borrowing it
Nothing to install: this file belongs to saajunaid/caddis-plugin. 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/saajunaid/caddis-plugin/main/.github/instructions/document-claims.instructions.mdgit clone --depth 1 https://github.com/saajunaid/caddis-pluginWrote 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/saajunaid/caddis-plugin/document-claims)<a href="https://agentmods.dev/instructions/saajunaid/caddis-plugin/document-claims"><img src="https://agentmods.dev/badge/instructions/saajunaid/caddis-plugin/document-claims/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/saajunaid/caddis-plugin/document-claims"><img src="https://agentmods.dev/badge/instructions/saajunaid/caddis-plugin/document-claims.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 | $0.00495 | $0.00495 |
| Opus 5.5 | $0.00198 | $0.00198 |
| Sonnet 5.5 | $0.00099 | $0.00099 |
| Haiku 4.5 | $0.00049 | $0.00049 |
Grade A, and why
caddis-plugin document-claims.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 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.
What it actually says
Claims in a document
Applies whenever you write or update a Markdown deliverable — a plan, a handover, a KB note, a PRD, a review, a parking-lot item, a README.
Before you write "probably", ask what the check would cost
A hedge is a cost transferred to whoever reads next. It is only worth transferring when the
check is genuinely expensive. One query, one grep, one file read — do it and write the answer.
Save the hedge for what actually cannot be settled from here, and then say why it cannot. "Unknown — the host is unreachable from this network" is useful. "Probably migration 011" is not.
Never record someone else's hedge as a fact
If a source says "most likely X", you may write "the previous session believed X, unverified" or you may go and check. You may not write "X".
A hedge recorded as a fact is worth less than no claim at all — it reads as settled, so nobody checks it again.
The incident that earned this
2026-09-08, two sessions handing work between them.
| Who | What they did |
|---|---|
| The writer | Said the dead Running rows were "most likely migration 011 landing" |
| The reader | Wrote "Migration 011 landed" into the relay doc |
Neither was true: those columns do not exist, and the mechanism is still unknown. The writer's own words afterwards — "a hedge is what you write when checking is expensive. It was not expensive — one query settled it, and I ran that query twenty minutes later anyway."
Both halves are faults. The writer's is the fixable one, because it removes the hazard instead of asking every later reader to handle it carefully.
What this is not
It is not a ban on uncertainty. It is a rule about cheap uncertainty. A document full of false confidence is worse than one that marks what it does not know — the point is to stop marking things unknown when finding out takes ten seconds.
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 First seen · 49 lines · 495 tokens per session scan A 19151c134e35
caddis-plugin document-claims.instructions.md is an instructions file published in the GitHub repository saajunaid/caddis-plugin (0 stars, last pushed today), licensed MIT. It adds 495 tokens to every session, about $0.0020 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 2026-09-11.
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.