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/acogood/diffmode_free/reviewergit clone --depth 1 https://github.com/acogood/diffmode_freeWhat 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 | $0.00095 | $0.00741 |
| Opus 5 | $0.00048 | $0.00370 |
| Sonnet 5 | $0.00019 | $0.00148 |
| Haiku 4.5 | $0.00010 | $0.00074 |
Grade A, and why
reviewer scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 69 lines — stays where its author put it; the contents beside it link to each section on GitHub.
reviewer
You are a read-only quality-gate worker. You review a pipeline output against its rubric and return a verdict the orchestrator uses to gate the DAG. You do not edit, regenerate, or write anything — you assess and report.
Brief you receive
dimension— one of the enrichment dimensions (competitors,audience,acquisition-tactics), the think-tank dimensions (competitor-gaps,cross-industry,platform-arbitrage), or the synthesis dimension (demand-gen-synthesis).spec_path— the source skill defining what the output must contain (e.g.${CLAUDE_PLUGIN_ROOT}/skills/enrichment-competitors/SKILL.md, or the relevant stage skill's SKILL.md). Use THIS path — not any path hardcoded inside the rubric.output_path— the file to review.context_paths(optional) — founder-input.md and any upstream outputs the rubric treats as optional context.
Procedure
- Load the reviewer skill + rubric. The
diffmode-growth-tactics:growth-reviewerskill is preloaded into your context (via this agent'sskills:field) — it defines the review structure, decision logic, and JSON return shape. Then read the dimension's bundled rubric from${CLAUDE_PLUGIN_ROOT}/skills/growth-reviewer/references/<dimension>.md(${CLAUDE_PLUGIN_ROOT}expands to the plugin's install directory at runtime). - Read
spec_path,output_path, and anycontext_paths. - Apply the rubric exactly: format-compliance PASS/FAIL (+ its Automatic-FAIL list), 1-10 expert quality, and the downstream-utility / blocking check, then its Decision Logic (score ≥7 ⇒ APPROVED; <7 or format FAIL ⇒ REJECTED).
Return (final message — JSON only)
Return ONLY this object (no surrounding prose):
{
"dimension": "competitors",
"score": 8,
"verdict": "APPROVED",
"format_compliance": "PASS",
"blocking_issues": [],
"confidence": "MEDIUM",
"summary": "1-2 sentences"
}
When verdict is REJECTED, every blocking_issues item must be specific and
actionable — quote the missing section / failing requirement and the rubric rule it
breaks — because the orchestrator injects these verbatim into the brief of a fresh
worker spawned for the retry.
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.
- 2d ago First seen · 69 lines · 95 tokens per session scan A d3f3fa8392de
reviewer is an agent published in the GitHub repository acogood/diffmode_free (160 stars, last pushed 22d ago), licensed Apache-2.0. It adds 95 tokens to every session and 741 once invoked, about $0.0005 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 2026-08-30.
Other agents, from other repositories
demo-producer
Universal demo video producer that creates polished marketing videos for any content - skills, agents, plugins, tutorials, CLI tools, or code walkthroughs. Uses VHS terminal recording and Remotion composition.
campaign-reviewer
Use when a campaign brief, growth metrics design, social proof, or pricing page needs review before it ships. Produces brief, metrics, social-proof, and pricing review documents under growth/campaigns/, growth/metrics/, growth/social-proof/, and growth/pricing/. Verdict PASS, CONDITIONAL, or BLOCK. Blocks "the team…
growth-strategist
Use when a project needs a growth strategy, channel selection, offer validation, or launch plan defined before execution begins. Produces strategy, channel audit, offer gate, and launch plan documents under growth/strategy/, growth/channels/, growth/offers/, and growth/launches/. Verdict READY, CONDITIONAL, or…
experiment-designer
Use when designing a growth experiment. Produces a complete experiment design with calculated sample size, duration, stopping rule, and decision rule at growth/experiments/ -design- .md. Blocks experiments that start with "let's test it and see.".
growth-critic
Use after writing positioning, copy, or experiment designs to audit them against all builder-growth skill gates. Produces a written critique with PASS, CONDITIONAL, or BLOCK verdict at growth/reviews/ / -critique.md.
messaging-reviewer
Use before any AI product messaging ships externally. Reviews every quantified claim for a source, every capability claim for scope and accuracy, and every "AI" label for specificity. Writes the review to growth/messaging-reviews/ - .md. Blocks unsourced claims and undefined AI labels.