Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/Chappygo-OS/Atomic-Specnpx agentmods add commands/chappygo-os/atomic-spec/clarifyWrote 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/chappygo-os/atomic-spec/clarify)<a href="https://agentmods.dev/commands/chappygo-os/atomic-spec/clarify"><img src="https://agentmods.dev/badge/commands/chappygo-os/atomic-spec/clarify/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/chappygo-os/atomic-spec/clarify"><img src="https://agentmods.dev/badge/commands/chappygo-os/atomic-spec/clarify.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 | $0.00065 | $0.05988 |
| Opus 5.5 | $0.00026 | $0.02395 |
| Sonnet 5.5 | $0.00013 | $0.01198 |
| Haiku 4.5 | $0.00006 | $0.00599 |
Grade A, and why
clarify 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 — 408 lines — stays where its author put it; the contents beside it link to each section on GitHub.
User Input
$ARGUMENTS
You MUST consider the user input before proceeding (if not empty).
Goal
Identify and resolve gaps that would cause the AI to make silent decisions during /atomicspec.plan or /atomicspec.implement. Two kinds of gaps:
- Spec ambiguity — things the user wrote but unclearly (v0.1 contract).
- Architectural lurkers + structural decisions — things the user didn't write but that the project needs answered before any code is generated. This is the v0.2 hardening (per Article IX Directive 7 scope).
Output: an updated spec.md (Clarifications section) AND new entries in specs/_defaults/registry.yaml with _provenance tags AND interview_completed: <today> set if the session completes normally.
This workflow runs BEFORE /atomicspec.plan. /atomicspec.plan Phase 0 will redirect users here when interview_completed: null (genesis behaviour). After it's set, /atomicspec.specify still advertises clarify (soft nudge) but does not block.
Six-phase flow
Phase 1: Mode question (Lite vs Detailed, sets cap)
Phase 2: Spec ambiguity scan (v0.1 contract — 11-category taxonomy)
Phase 3: Architectural lurkers (from .specify/knowledge/architectural-lurkers.yaml)
Phase 4: Trigger-driven probes (from .specify/knowledge/triggers.yaml)
Phase 5: Compliance probes (compliance:true triggers, two-step gate)
Phase 6: Write + provenance (batched registry write, accept-rate audit)
Registry Protocol (Constitution Directive 7)
Follow _registry-protocol.md:
- On entry: Read
specs/_defaults/registry.yaml. Load existing values so we don't re-ask fields that are already non-null. - During: Track every accepted answer in working memory. Mark each as
human(active choice) oraccepted_recommendation(default taken) for provenance tagging. - On exit: Batch-write to
registry.yaml. Setinterview_completedif the session completed normally. Update_provenanceblock with one entry per write.
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 · 408 lines · 65 tokens per session scan A e60894980588
clarify is a command published in the GitHub repository Chappygo-OS/Atomic-Spec (10 stars, last pushed 1mo ago), licensed MIT. It adds 65 tokens to every session and 5,988 once invoked, about $0.0003 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
ijfw-audit
Run the IJFW audit gate for the current workflow phase. Usage: /ijfw-audit [phase name].
constitution
A command for creating or updating a project's constitution, a document that records its governing principles and working rules, in Simplified Chinese.
ijfw-execute
Jump directly to the IJFW workflow Execute phase (Deep D4 / Quick Q3). Usage: /ijfw-execute [task or phase name].
ijfw-metrics
Show IJFW token / cost / session metrics over a time window. Reads .ijfw/metrics/sessions.jsonl via the ijfwmetrics MCP tool.
assistant-auto
Orchestrator in automatic mode. Choose the workflow that semantically fits based on the request + the injected repo context, then execute immediately via Skill.
doc-changelog
Generation and maintenance of the project changelog.