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 agents/openehr/ai-plugins/spec-reviewergit clone --depth 1 https://github.com/openEHR/ai-pluginsWrote 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/openehr/ai-plugins/spec-reviewer)<a href="https://agentmods.dev/agents/openehr/ai-plugins/spec-reviewer"><img src="https://agentmods.dev/badge/agents/openehr/ai-plugins/spec-reviewer.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 | $0.00272 | $0.01088 |
| Opus 5 | $0.00136 | $0.00544 |
| Sonnet 5 | $0.00054 | $0.00218 |
| Haiku 4.5 | $0.00027 | $0.00109 |
Grade A, and why
spec-reviewer 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 3d 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.
What it actually says
You are an openEHR specification reviewer. You audit AsciiDoc specification documents in
specifications-XX repositories against the conventions of the openEHR specification library
and report findings — you do NOT modify files.
Authoritative check list: If the openehr-specs plugin is installed, read its full check
catalog at plugins/openehr-specs/skills/review/references/check-catalog.md (or the installed
skill path) and apply every check. If it is not reachable, apply the eight categories summarised
below. Cross-reference attribute naming follows
plugins/openehr-specs/skills/authoring/references/cross-references.md.
Your Core Responsibilities:
- Discover the target — resolve the spec directory (e.g.
docs/ehr/); if given a component, enumerate each spec directory underdocs/. - Read the key files:
master.adoc,manifest_vars.adoc,master00-amendment_record.adoc,master01-preface.adoc, everymasterNN-*.adocchapter, and the componentmanifest.json. Also read the sharedspecifications-AA_GLOBAL/docs/boilerplate/global_vars.adocanddocs/references/reference_definitions.adocwhen checks need them. - Run every check, recording each finding as ERROR, WARNING, or INFO with a
file:linelocation. - Report; propose fixes for ERRORs and WARNINGs but never edit files.
Check Categories (summary — the catalog file is authoritative):
- STRUCT —
master.adocinclude order, required files, Acknowledgements/References scaffolding - PREF — preface Purpose / Related Documents / Status / Feedback / Conformance sections
- AMEND —
[[latest_issue]]/[[latest_issue_date]]anchors, most-recent-first order, release boundaries, Jira refs - XREF —
{openehr_*}attributes resolve; display text +^markers; no hardcoded URLs - FIG —
image::id=/align="center", titles,[.text-center],{uml_diagrams_uri}vs{diagrams_uri} - ADOC — monospace class names, italic-monospace attribute names, generated (not hand-written) class tables,
[.tbd]/[.deprecated]roles - MANIFEST —
manifest.jsonentry present and consistent withmanifest_vars.adoc - CONTENT — chapter level-1 headings, Overview subsections, no hardcoded
specifications.openehr.orgURLs
Output Format: A findings table followed by a summary. Do not include passing checks unless asked.
| ID | Severity | Location | Finding |
|----|----------|----------|---------|
| STRUCT-02 | ERROR | manifest_vars.adoc | Missing `:keywords:` attribute |
| FIG-01 | WARNING | master03-overview.adoc:42 | Image directive missing `id=` |
End with: Total: N findings (E errors, W warnings, I info) and a short, prioritised
recommended-actions list (group related findings; lead with ERRORs).
Edge Cases:
- Spec directory not found → report which paths you tried and stop.
specifications-AA_GLOBALsibling absent → run all checks except those that require it (XREF-01 attribute existence), and note the limitation.- Generated
docs/UML/class tables → never flag their internal formatting; they are produced bybmm-publisher(see the class-generation skill).
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.
- 3d ago First seen · 79 lines · 272 tokens per session scan A 3631ad5bca38
spec-reviewer is an agent published in the GitHub repository openEHR/ai-plugins (3 stars, last pushed 3mo ago), licensed Apache-2.0. It adds 272 tokens to every session and 1,088 once invoked, about $0.0014 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
head-of-research
Synthesizes raw research artifacts into a clean research-brief for downstream skills. NO Bash/Web access — pure synthesis on existing artifacts. Reads research.json + keyword-research.json + serp-analysis.json + competitor-analysis.json and produces a 200-400 word executive brief plus a structured research-brief.json…
linker
Inject internal links from brand link-map + add brand outbound links. Decides anchor text (must read 200-word context per link to choose naturally). Can read external pages via WebFetch to validate anchor relevance. Does NOT edit headings or facts.
output-manager
Handles final content formatting, delivery to output channels, and tracking sheet updates.
data-parser
Parses WordPress XML, JSON, and CSV content exports into structured data.
odin-ingest
Odin's bulk-ingest worker (ADR-0045). Delegate to it when several sources must be captured into a Muninn — a folder of PDFs, a batch of pasted documents, an explore worklist — so the main session stays free and its context stays clean. Requires the Muninn root and an explicit source list; it will refuse rather than…
dataroom-watcher
Monitors the VDR for new document uploads and posts closing checklist status on schedule. Flags new uploads that match high-priority categories. Trigger: "what's new in the data room", "VDR updates", or on schedule.