agentmods.dev

evidence_prosecutor

evidence_prosecutor is an agent for coding agents from antonio0720/writing-intelligence. It costs 0 tokens per session (396 once invoked), scanned A, original, MIT.

An agent for checking whether written claims are supported by available evidence. It records the status of claims, identifies risky wording, and blocks made-up citations, statistics, quotes, or studies.

In plain words
What is it for?
Use it to review reports or other evidence-based writing, classify claims, check sources, flag words such as “always” or “must,” and record decisions about questionable statements.
Why use it?
It helps prevent unsupported statements and overconfident language from reaching the final output. It also makes uncertain claims visible so they can be cited, softened, qualified, or removed.

Agent

not rated 13 26d ago A scan 0 tokens original MIT
View the source file antonio0720/writing-intelligence
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/antonio0720/writing-intelligence/evidence_prosecutor
Clone the repo
git clone --depth 1 https://github.com/antonio0720/writing-intelligence

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 evidence_prosecutor

README.md
[![agentmods](https://agentmods.dev/badge/agents/antonio0720/writing-intelligence/evidence_prosecutor.svg)](https://agentmods.dev/agents/antonio0720/writing-intelligence/evidence_prosecutor)
Your own site
<a href="https://agentmods.dev/agents/antonio0720/writing-intelligence/evidence_prosecutor"><img src="https://agentmods.dev/badge/agents/antonio0720/writing-intelligence/evidence_prosecutor.svg" alt="Measured on agentmods" height="20"></a>
Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 396 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 $0.00000 $0.00396
Opus 5 $0.00000 $0.00198
Sonnet 5 $0.00000 $0.00079
Haiku 4.5 $0.00000 $0.00040

Measured 3d ago against content hash 737a8bc63298, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

evidence_prosecutor 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.

agents/evidence_prosecutor.md · 45 lines

What it actually says

Evidence Prosecutor

Pass: 5 Artifact: EpistemicLedgerV3 (schemas/epistemic_ledger.schema.json) Doctrine: references/compiler/epistemic_ledger.md

Job

Classify every major claim. Mark source status. Block fabrication. Surface universal-quantifier and inflated-verb risk. Decide on each flagged claim: cite, soften, qualify, remove, or escalate.

Inputs

  • Intake contract (Pass 0)
  • Corpus map (Pass 2)
  • Architecture graph (Pass 4)

Outputs

  • An EpistemicLedgerV3 with claims classified
  • A decisions log
  • A delivery_block flag

Behavior

  1. Walk every claim node in the architecture graph plus every sentence not bound to a node but making an assertion.
  2. Classify each as observed_fact / sourced_fact / inference / synthesis / recommendation / rhetoric.
  3. Assign source_status per the corpus map.
  4. Flag universal quantifiers (all, every, none, never, always, must).
  5. Flag inflated verbs (revolutionize, transform, redefine, disrupt).
  6. Detect fabrication risk: invented citation, statistic, quote, study, etc.
  7. For each flagged claim, pick a decision (cite_added / claim_softened / claim_qualified / claim_removed / user_clarification_required).
  8. If any fabrication remains blocked or any source is unsafe, set delivery_block: true.

Hard Rules

  • Mandatory in high-stakes domains; cannot be skipped.
  • Never lower a fabrication block silently.
  • Scorekeeper cannot override the block.
  • Every decision must have a rationale.

Hands Off To

  • Sentence Surgeon (Pass 6) — to implement softening/qualification
  • Scorekeeper (Pass 10) — for the Epistemic Integrity score
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. 3d ago First seen · 45 lines · 0 tokens per session scan A 737a8bc63298

Subscribe to this mod's changes

evidence_prosecutor is an agent published in the GitHub repository antonio0720/writing-intelligence (13 stars, last pushed 26d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 396 tokens. 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.

Researchanti-slopclaude-skillscreative-writinggrant-writingscreenwritingwriting-assistant
Related

Other agents, from other repositories

chapter-extractor

章节摘要与情节点提取专家。接收单章文本，输出结构化摘要、情节点列表、角色提及。 被 story-long-analyze（拆解管道 Stage 2）按章节并行调用。 输出格式严格遵循本文件「输出格式」章节；不依赖外部输出模板文件。.

zenstory-ai/oh-story-claudecode · 70 tokens

story-explorer

故事项目结构化查询 agent（只读）。响应关于角色状态、伏笔进度、设定出现位置、 时间线节点、写作进度的查询。使用 grep + read 从项目文件系统中检索信息， 返回结构化 JSON 摘要。 被 story-long-write（日更 Step 1 上下文加载）、story-review（审查时查设定）、 story 路由（用户自然提问时）调用。 不做任何创作判断或修改。.

zenstory-ai/oh-story-claudecode · 104 tokens

story-architect

故事架构与世界观创作专家。负责题材选择、核心梗设计、世界观构建、大纲排布、 钩子/悬念/反转等叙事工程、情绪弧线设计、范围控制审查。 被 story-long-write（Phase 1-3）、story-short-write（Phase 1-2）调用。 也可审查已有内容的结构问题。.

zenstory-ai/oh-story-claudecode · 90 tokens

humanizer

识别并修复典型 AI 写作特征，从内容、语言、风格三个维度进行"净化"，并强化原稿中已有的观点、节奏、不确定性和个人视角。Humanizer 只改表达，不创造事实、经历或证据。包含严格的黑名单过滤和 50 分制质量自评。.

dongbeixiaohuo/writing-agent · 65 tokens

empathy-designer

社交货币与共情设计师。根据大纲和伤疤细节，设计文章的分享动因（Impression Management），建立 Share Map。由工作流导演在 Stage 4 显式调用。.

dongbeixiaohuo/writing-agent · 53 tokens

opening-tournament

开头赛马机制。根据前序策划信息，并行提供 3 种极具差异化的实战开头原型方案（字数 150-300 字/个），由工作流导演在 Stage 5.8 显式调用。.

dongbeixiaohuo/writing-agent · 62 tokens