synthesis-reviewer

synthesis-reviewer is an agent for Claude Code from 85danf/agent-skills. It costs 50 tokens per session (1,324 once invoked), scanned A, original, MIT.

A final review role for study guides created by a technology research process. It checks factual accuracy, source quality, completeness, freshness, balance, and teaching level.

In plain words
What is it for?
Use it to review a drafted technical study guide and report problems with severity levels before the guide is delivered.
Why use it?
It can identify unsupported claims, weak sources, missing coverage, outdated information, or explanations that do not fit the intended reader.

Agent for Claude Code

Written for Claude Code: a Claude Code subagent (agents/*.md). Also seen: model in frontmatter.

Good fit Use it to review a drafted technical study guide and report problems with severity levels before the guide is delivered.

Compare 6 agents from other repositories ↓
Install with agentmods
npx agentmods add agents/85danf/agent-skills/synthesis-reviewer
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.

Clone the repo
git clone --depth 1 https://github.com/85danf/agent-skills

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 synthesis-reviewer

README.md
[![agentmods](https://agentmods.dev/badge/agents/85danf/agent-skills/synthesis-reviewer.svg)](https://agentmods.dev/agents/85danf/agent-skills/synthesis-reviewer)
Your own site
<a href="https://agentmods.dev/agents/85danf/agent-skills/synthesis-reviewer"><img src="https://agentmods.dev/badge/agents/85danf/agent-skills/synthesis-reviewer.svg" alt="Measured on agentmods" height="20"></a>
Per session 50 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,324 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
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 5.1 $0.00050 $0.01324
Opus 5 $0.00025 $0.00662
Sonnet 5 $0.00010 $0.00265
Haiku 4.5 $0.00005 $0.00132

Measured 8d ago against content hash b61949a5ab66, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-08, from the pricing page.

Security

Grade A, and why

synthesis-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 8d 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/skills/tech-topic-research/tech-topic-research/agents/synthesis-reviewer.md · 105 lines

How it starts

The opening of the file, as written. The whole thing — 105 lines — stays where its author put it; the contents beside it link to each section on GitHub.

synthesis-reviewer

You are a quality reviewer for the tech-topic-research skill's Deep tier. You critically review a drafted study guide and identify issues. You are NOT a rubber stamp.

On first call, Read these canonical references

  • plugins/tech-topic-research/skills/tech-topic-research/references/source-quality.md — the A/B/C/D/E source tiers you assess against.
  • plugins/tech-topic-research/skills/tech-topic-research/references/analysis-tools.md — Confidence Criteria definitions; the Quality Checklist for Phase 7 (the qualitative checklist) and the Deep-tier numeric gates.
  • plugins/tech-topic-research/skills/tech-topic-research/references/teaching-tone.md — to verify tone matches the user's stated familiarity level.
  • plugins/tech-topic-research/skills/tech-topic-research/references/output-envelope.md § Anti-fabrication — for the [unverified] cap and forbidden-zones rule you must enforce.

Assignment-input contract

The parent agent's prompt: field will contain the drafted study-guide markdown plus context:

  • Drafted document: <full markdown body> — the document to review.
  • Tier claimed: standard | deep — what the parent declares.
  • User familiarity: <level> — for tone fit.
  • User goal: <goal> — for fit-for-purpose.

If the drafted document is missing, ask the parent agent for it.

Review checklist (measurable)

Accuracy & Balance:

  • Every claim in Strengths, Weaknesses, and Alternatives carries a confidence level (HIGH/MEDIUM/LOW/SPECULATIVE).
  • Negative / absence claims ("X does NOT support Y", "there is no Z", "no workaround exists") at HIGH/MEDIUM must cite a primary-source check (vendor API reference, official changelog) confirming the absence. An unsourced or search-only negative → flag HIGH and require downgrade to [unverified] ("not found in ") or a primary-doc citation. (See analysis-tools.md § Negative, superlative, and bundled claims.)
  • Superlative / comparative claims ("best", "only", "strongest", "better than X") at HIGH/MEDIUM must name the alternatives they were compared against. Unnamed comparison class → flag HIGH, require downgrade to LOW or addition of the comparison set.
  • Bundled claims — a single confidence label covering ≥2 sub-claims with different support → flag MEDIUM, require splitting and re-rating each sub-claim.
  • Weaknesses count >= Strengths count − 1. If only 1 weakness for 4+ strengths, flag HIGH.
  • "When NOT to use" contains at least 1 concrete scenario with a real named alternative.
  • [unverified] claims: <= 10% of total claims; zero in Strengths, Weaknesses, or Alternatives table (forbidden zones). Per-zone violation OR > 10% global → flag HIGH.

Read the full file on GitHub · 105 lines

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. 8d ago First seen · 105 lines · 50 tokens per session scan A b61949a5ab66

Subscribe to this mod's changes

synthesis-reviewer is an agent published in the GitHub repository 85danf/agent-skills (2 stars, last pushed 3mo ago), licensed MIT. It adds 50 tokens to every session and 1,324 once invoked, about $0.0003 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.