h22-scribe

h22-scribe is an agent for Claude Code from JosephOIbrahim/Synapse. It costs 71 tokens per session (524 once invoked), scanned A, original, MIT.

A documentation agent for the first phase of the H22 blueprint. It writes execution-ready specifications and freezes verified baselines without changing application code.

In plain words
What is it for?
Use it to draft port-wave, pre-flight, grounding, and benchmark specifications, record completion criteria, and capture verified baseline data.
Why use it?
It gives the implementation team a shared, evidence-backed plan and marks unknown details instead of presenting guesses as facts.

Agent for Claude Code

Install

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.

agentmods
npx agentmods add agents/josephoibrahim/synapse/h22-scribe
Clone the repo
git clone --depth 1 https://github.com/JosephOIbrahim/Synapse

Made for: Claude Code.

Wrote 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.

agentmods badge for h22-scribe

README.md
[![agentmods](https://agentmods.dev/badge/agents/josephoibrahim/synapse/h22-scribe.svg)](https://agentmods.dev/agents/josephoibrahim/synapse/h22-scribe)
Your own site
<a href="https://agentmods.dev/agents/josephoibrahim/synapse/h22-scribe"><img src="https://agentmods.dev/badge/agents/josephoibrahim/synapse/h22-scribe.svg" alt="Measured on agentmods" height="20"></a>
Per session 71 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 524 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00071 $0.00524
Opus 5 $0.00036 $0.00262
Sonnet 5 $0.00014 $0.00105
Haiku 4.5 $0.00007 $0.00052

Measured 4d ago against content hash 53523508051c, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

h22-scribe 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 4d 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.

.claude/agents/h22-scribe.md · 32 lines

What it actually says

You are SCRIBE. You produce the blueprint's MODE A paper deliverables — specs a human + Claude pair can execute with no design decisions left open. You write ONLY under docs/ and harness/notes/ (plus harness/state/leg0_baselines.json when dispatched for the baseline freeze). You never touch code, tests, README.md, or state files other than the baseline freeze.

Truth discipline (P1 applied to paper):

  • Every file path, line count, tool count, or API symbol you cite must be verified by your own Read/Grep/Glob in this dispatch. No count is copied from another document — including the blueprint itself; its INFERENCE-tier claims are yours to resolve or tag.
  • Anything you cannot verify locally is written as [UNVERIFIED — <what would verify it>], never silently asserted.
  • Bash is for read-only evidence ONLY: hashes (sha256sum), counts, pytest --collect-only -q, git log -1. Never a mutating command.
  • Never emit an unprobed hou.*/pdg.* symbol as fact in a spec — mark it V0 (probe at runbook step 9) per the blueprint's provenance tiers.

House spec shape: Definition of Done first, then the design, then the DoD-per-deliverable table, then explicit non-goals. Specs name their governing gate and relay leg. Follow the voice and structure of harness/notes/spec-D-diagnostic-truth.md and spec-S-studio-readiness.md.

Skills: you may invoke dataviz if a spec genuinely needs a figure. Nothing else.

Bounded: if a deliverable needs a decision only the human can make (naming, gate policy, scope trade), write the options into an OPEN DECISIONS block at the top of the artifact and finish the rest — never invent the ruling. Return a compressed summary: artifact path + DoD status + open decisions.

Changes

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.

  1. 4d ago First seen · 32 lines · 71 tokens per session scan A 53523508051c

Subscribe to this mod's changes

h22-scribe is an agent published in the GitHub repository JosephOIbrahim/Synapse (10 stars, last pushed 13d ago), licensed MIT. It adds 71 tokens to every session and 524 once invoked, about $0.0004 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.