Getting it into your agent
This one installs as part of its plugin. Adding the marketplace and installing the plugin brings it with everything else the plugin ships.
/plugin marketplace add Linaqruf/kana-code-plugins/plugin install kana-specWrote 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/commands/linaqruf/kana-code-plugins/kana-spec)<a href="https://agentmods.dev/commands/linaqruf/kana-code-plugins/kana-spec"><img src="https://agentmods.dev/badge/commands/linaqruf/kana-code-plugins/kana-spec/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/commands/linaqruf/kana-code-plugins/kana-spec"><img src="https://agentmods.dev/badge/commands/linaqruf/kana-code-plugins/kana-spec.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 | $0.00112 | $0.01963 |
| Opus 5.5 | $0.00045 | $0.00785 |
| Sonnet 5.5 | $0.00022 | $0.00393 |
| Haiku 4.5 | $0.00011 | $0.00196 |
Grade A, and why
kana-spec 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 — 162 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Subject of this spec engagement: $ARGUMENTS (free-form — infer; if empty, the subject is the whole project)
The Job
The product is not the spec file — it is better future agent sessions. A
spec front-loads decisions into durable contracts: the spec is what a future
session READS, SPEC/ files are what it LOOKS UP, CLAUDE.md points at both,
prompt.md restarts the loop. Every choice below serves that: a fresh session
with zero context must be able to act correctly from the artifacts alone.
The Model: Subject × Mode
Every engagement is one pipeline over two coordinates:
- Subject — what is being specified: the whole project, a feature, the
design system, an API, a migration. Infer from the user's words + scout
findings. Scale follows: project-scale →
SPEC.mdat root; component-scale →SPEC/<slug>.md. - Mode — Plan (subject doesn't exist yet), Document (it exists; capture reality), Overhaul (it exists; audit it, then re-derive from first principles with a phased migration checklist).
There are no other routes. Legacy-style arguments ("feature auth", "design:overhaul") read naturally as subject/mode hints.
Pipeline
Steps are tagged: [execute] = do directly, no deliberation; [reason deeply] = slow down, this is where judgment earns its keep.
1. SCOUT [execute]
ALWAYS spawn the spec-scout agent before framing — skipping it is only
allowed for true greenfield (empty directory). Pass: the subject + repo path.
If the repo is large (>~200 source files) or the subject spans clearly
separate concerns, spawn 2-3 scouts in parallel in the same message, each
with a narrower beat (e.g. structure+stack / data+API / conventions+tests).
The scout returns a cited map: every claim carries file:line evidence or is
marked unknown. Treat it as the interview's pre-fill source, never as
something to re-derive yourself in the main context.
2. FRAME [execute]
One AskUserQuestion confirming subject, mode, and scale — options pre-filled from scout findings (e.g. if the scout found an existing SPEC.md, offer Update vs Fresh; if the user asked "what next?" with no subject, present the gap between scout map and existing spec as candidate subjects).
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 · 162 lines · 112 tokens per session scan A b0dddb148493
kana-spec is a command published in the GitHub repository Linaqruf/kana-code-plugins (4 stars, last pushed 22d ago), licensed MIT. It adds 112 tokens to every session and 1,963 once invoked, about $0.0004 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 commands, from other repositories
permissions
Create or update DOGMA-PERMISSIONS.md to configure what Claude can do autonomously.
merge
You are executing the /hydra:merge command. Merge a worktree branch back into the current branch.
watch
You are executing the /hydra:watch command. Live monitoring of all background agents with real-time status updates.
update
You are executing the /import:update command. Re-fetch documentation from original sources.
spawn
You are executing the /hydra:spawn command. Start an agent in an existing worktree.
sandbox-promote
Promote an accepted, version-worthy WIP artifact from .credo/sandbox-tmp/ / to .credo/sandbox/ /. is the sandbox-tmp folder name (argument: $1).