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/mininglamp-oss/octo-cli/claude-mdgit clone --depth 1 https://github.com/Mininglamp-OSS/octo-cliWrote 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/mininglamp-oss/octo-cli/claude-md)<a href="https://agentmods.dev/instructions/mininglamp-oss/octo-cli/claude-md"><img src="https://agentmods.dev/badge/instructions/mininglamp-oss/octo-cli/claude-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 | $0.04507 | $0.04507 |
| Opus 5 | $0.02253 | $0.02253 |
| Sonnet 5 | $0.00901 | $0.00901 |
| Haiku 4.5 | $0.00451 | $0.00451 |
Grade A, and why
octo-cli CLAUDE.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 — 201 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md — octo-cli project instructions
What is octo-cli
octo-cli is the command-line interface for the Octo ecosystem, built for AI Agent Bots to call via exec from agent runtimes (OpenClaw, Claude Code, etc.). Output is a JSON envelope; agent-runtime commands take no interactive I/O. The sole exception is octo-cli auth login, an operator-only setup step that reads the token from a hidden terminal prompt (or stdin).
Architecture
- Go single binary, cobra CLI framework.
- Metadata-driven: the entire service command tree is auto-registered at startup from OpenAPI 3.x specs embedded into the binary via
internal/registry. To add or change an endpoint, update a spec — not code. Spec extensions cover identity routing (x-octo-allowed-token-kinds,x-octo-mount-by-token-kind), output shaping (x-octo-response-fields,x-octo-lossless-id-fields), and log hygiene (x-octo-secret); each is opt-in, and omitting it keeps the historical behaviour byte-for-byte.x-octo-flagrenames the flag for a query/header param or a promoted body field, and on a path param it adds an optional flag alternative to the positional slot (the escape hatch for base64url ids starting with-, which cobra would parse as a flag) — an operation with no such declaration keepscobra.ExactArgs. - Thin client: all business logic lives in backend services (matters, dmworkim). CLI is transport + validation + formatting.
- Unified gateway: every domain uses
OCTO_API_BASE_URL; embedded OpenAPI paths carry module namespaces such as/market/api,/docs-html, and/fleet/api. - Factory DI:
internal/cmdutil.Factoryis the DI container; no package-level globals. Tests inject stubs throughConfigFunc/CredentialFunc/ClientFunc/RegistryFunc.Factory.ErrorEmittedtracks whether an error envelope was already written to stderr, preventing double-emit between RunE and the top-level main error handler. - JSON envelope I/O:
{ok, identity, data, _pagination, _rate_limit}on stdout for success;{ok:false, error:{type,code,message,hint,detail}}on stderr for failure. Exit codes: auth=3, validation/config=2, rest=1. - Pre-flight validation (
cmd/service/run.go,cmd/service/enum.go): the metadata-driven path checks the resolved request against the spec before any HTTP — required fields andminItems(VALIDATION_ERROR), specenumvocabularies (ENUM_NOT_ALLOWED),uint64id range, and a refusal of empty or./..path values (an empty value addresses the collection and a gateway that normalises dot segments retargets the request — and the engine emitsDELETE). Loop Public API operations and services that opt intox-octo-strict-request-schema(currently Marketplace) additionally enforce composition, closed-object and minimum-property constraints, Unicode string lengths and maximum array sizes; those extended checks remain backend-enforced for other legacy domains. Body checks walk the merged body, so--datais validated too, at every nesting depth and includingformat: uint64—--datais not a raw passthrough on this path, and it decodes withUseNumberso an id above 2^53 is never rounded before it is checked. A property present with an explicitnullis checked, not skipped: it is refused wherever the schema constrains the value (an enum or auint64format), becausenullmatches no vocabulary member and decodes into a scalar field as the zero value — for aparent_idthat is folder0, the documented root, i.e. a valid id addressing a place nobody named. A required field set tonullis reported as missing instead, so both paths refuse it; anullon an unconstrained property is still forwarded, since a backend may accept it to clear the field. Enum and constant comparison is by canonical form, not==, because the same wire value arrives asint,float64orjson.Numberdepending on how it was supplied; numbers compare by exact decimal text, so an integer vocabulary admits1but not1.0— the body keeps what the caller wrote, and a non-integer would fail at the backend on a value the local gate had called valid. Hand-written composites reuse the same walker viaservice.ValidateRequestBody, so a composite that replaces a generated leaf cannot enforce a weaker contract than the leaf did.
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 · +11 lines · +105 tokens per session 536b67129b7a
- 5d ago First seen · 190 lines · 4,402 tokens per session scan A 9ef15ce92bec
octo-cli CLAUDE.md is an instructions file published in the GitHub repository Mininglamp-OSS/octo-cli (673 stars, last pushed today), licensed Apache-2.0. It adds 4,507 tokens to every session, about $0.0225 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
Alita_Robot AGENTS.md
AGENTS.md instructions for Divkix/Alita_Robot, covering repository guidelines, 0. maintaining this file, 1. mental model, 2. project structure and 3. build, test, lint.
imbot-sdk-go CLAUDE.md
Instructions for juggleim/imbot-sdk-go, covering claude.md, commands, architecture, transport core (imbotclients/imbotclient.go) and feature methods (one file per domain in imbotclients/).
Alita_Robot CLAUDE.md
Claude Code instructions for Divkix/Alita_Robot, a project described as: 🤖 Alita is a Telegram Group management bot made using Gotgbot and Go, which makes it modern and ⏩ faster than most of the Telegram chat managers. 💬.
Alita_Robot GEMINI.md
Gemini CLI instructions for Divkix/Alita_Robot, a project described as: 🤖 Alita is a Telegram Group management bot made using Gotgbot and Go, which makes it modern and ⏩ faster than most of the Telegram chat managers. 💬.
ollama CLAUDE.md
Claude Code instructions for ollama/ollama: See AGENTS.md for the shared agent instructions for this repository.
dingtalk-workspace-cli AGENTS.md
AGENTS.md instructions for DingTalk-Real-AI/dingtalk-workspace-cli, covering repository agent guide, build and test, command framework declaration, flag / help / schema homology and agent schema contract.