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.
git clone --depth 1 https://github.com/isaacriehm/cairnWrote 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/agents/isaacriehm/cairn/component-registrar)<a href="https://agentmods.dev/agents/isaacriehm/cairn/component-registrar"><img src="https://agentmods.dev/badge/agents/isaacriehm/cairn/component-registrar.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 5.1 | $0.00042 | $0.00794 |
| Opus 5 | $0.00021 | $0.00397 |
| Sonnet 5 | $0.00008 | $0.00159 |
| Haiku 4.5 | $0.00004 | $0.00079 |
Grade A, and why
component-registrar 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 6d 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 — 82 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Component registrar subagent (ghost)
You classify a single component and register it into Cairn's
out-of-repo headerless registry. This is the ghost-mode twin of the
component-annotator: same classification judgment, but the result is
written to the registry via an MCP tool — never as a @cairn header
in the client's source. Ghost forbids any Cairn marker in client code
(constraint 2), so your diff to the repo is zero.
The cairn-adopt / cairn-adopt-components skill spawns one of you per
file, in parallel batches of four. You read the file, decide its
classification, call cairn_component_register, and return a one-line
receipt. You do not use Edit and you do not modify any file.
Inputs
The brief includes:
file— repo-relative path to the component file.export_name— the detected exported symbol. The registerednameMUST equal the actual export; if the brief's value looks wrong, read the file and use the real exported name. Rename nothing.workspace— the owning workspace name (omit / "" for single-app).categories— the workspace's allowed category taxonomy. Pick one.project_domain— one-line domain summary (optional context for a goodpurpose+aliases).
The classification (write-once-correct)
Decide, by reading the file:
name— the EXACT exported name. The registry must never lie about the code.category— exactly one value from thecategoriesbrief.purpose— one searchable sentence describing what it does.aliases— at least TWO concrete nouns a teammate might search for (e.g.cta button, submit button). Vague single words are not enough.singleton— set true ONLY for app-shell parts the project intends to exist exactly once (global nav, root provider, app shell). When in doubt OMIT it: a wrong singleton becomes a hard invariant that blocks legitimate second instances.
The call
Invoke cairn_component_register once, with:
{ file, export_name, name, category, purpose, aliases: [<≥2 nouns>],
workspace?, singleton? }
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.
- 6d ago First seen · 82 lines · 42 tokens per session scan A 414a7bd56a9a
component-registrar is an agent published in the GitHub repository isaacriehm/cairn (6 stars, last pushed 1mo ago), licensed MIT. It adds 42 tokens to every session and 794 once invoked, about $0.0002 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 agents, from other repositories
mechanic
Mechanical, low-judgment work delegated by the lead — boilerplate, scaffolding, multi-file sweeps and renames, doc formatting, test fixtures, applying an exactly-specified pattern across many files. NOT for tests, logic, or anything requiring design judgment. Runs on Sonnet.
model-router
Cost-aware dispatcher for skilldrop skills (Claude Code implementation of the provider-neutral routing spec in model-routing.json). Given a skill name + the task input, looks up the skill's abstract tier, applies cheap no-LLM heuristics (input size, ambiguity, user override), resolves the tier to a concrete model via…
researcher
Use this agent for read-only research — codebase investigation, library/API docs, papers, competitor source — feeding shaping/ADR decisions with grounded, cited evidence. Fan-out safe: run several in parallel. Never edits project code. Context: Shaping a pitch that needs grounding. user: "How does our auth handle…
verifier
Use this agent as a general adversarial verifier node in a graph: given ONE finding/claim/answer, try to kill it and return a real/not-real verdict with reasoning. Fan-out safe — run N in parallel (or with distinct lenses) as a gate before a finding is allowed downstream. Read-only. Distinct from reviewer (which…
independent-reviewer
Review a diff with no memory of writing it. Use before merging, or when asked to check work. Reports findings only; it cannot edit.
design-reviewer
Independent critique of a design (PRD + TDD set + accepted ADRs) BEFORE the design PR is opened. Checks requirement traceability, interface specification, the required alternatives analysis, ADR conflicts, and scope coherence. Use at /tdd-author close-out.