Skip to main content
skill CX Codex

evidence_synthesis — Claude Code skill for evidence synthesis

The evidence_synthesis Claude Code skill builds concise answers from supplied observations and task materials. It links claims to evidence and distinguishes reported facts from model inference.

no license ★ 776 updated 2026-11-12

Install into Claude Code · Manually: put the files in place All methods and platforms ↓
~/.claude/skills/evidence_synthesis/SKILL.md
Open on GitHub

What evidence_synthesis does

The skill turns supplied observations into a concise response and keeps a compact map connecting claims, supporting evidence, and URLs. It distinguishes source facts from arithmetic results and model inference, then follows the requested answer format. Brief citations can accompany the response when allowed.

For arithmetic, it can use tool.python_script_runner.v1, preserving units and the origin of input values. The skill itself makes no network requests; a separate tool must retrieve source material. Its work starts with information already supplied.

What it helps with

  • condense observations into a short answer
  • connect claims with evidence and URLs
  • separate calculated results from model inference
  • preserve units and the origin of numerical inputs

Example prompts for the agent

  • Write a concise answer from these observations and attach supporting evidence.
  • Separate source facts, calculations, and model inference in the response.
  • Create a claim-to-evidence-and-URL map from the supplied material.

How to install evidence_synthesis

Manually: put the files in place

~/.claude/skills/evidence_synthesis/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/evidence_synthesis/

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about evidence_synthesis

How do I install evidence_synthesis in Claude Code?

Create the skill file at ~/.claude/skills/evidence_synthesis/SKILL.md and place the skill contents there.

What does evidence_synthesis do?

It drafts concise answers from supplied material, connects claims to evidence, and keeps source facts, calculations, and model inference distinct.

Does evidence_synthesis access the internet?

No. The skill does not make network requests; source retrieval is handled through a separate tool operation.

Similar skills

All skills for Science and research →

From the same set24

The whole set →
test_driven_qa test_driven_qa — a subagent for Claude Code turns requirements and repository materials into a QA plan or test artifact. It ranks test scenarios and keeps observed evidence separate from proposed checks. CX Codex ★ 776repo stars performance_engineer The performance_engineer Claude Code subagent finds application, API, database, and infrastructure bottlenecks from measurements. It builds an optimization plan with baselines, experiments, and regression checks. CX Codex ★ 776repo stars refactoring_specialist refactoring_specialist — a subagent for Claude Code plans code refactoring while preserving observable behavior. It pairs structural changes with regression protection. CX Codex ★ 776repo stars research_queries The research_queries Claude Code skill turns research questions into targeted searches. It identifies entities, dates and constraints, then looks for evidence behind unresolved claims. CX Codex +1 ★ 776repo stars source_verification The source_verification Claude Code skill ties claims to supporting sources. It helps check dates, figures, and entity identity in research responses. CX Codex +1 ★ 776repo stars web_navigation The web_navigation Claude Code skill reads public web pages and gathers evidence for specific facts. It follows relevant links to find supporting details. CX Codex +1 ★ 776repo stars frontend_developer The frontend_developer Claude Code subagent turns UI requirements and repository context into an implementation plan and code-ready changes. It covers responsive layouts, accessibility, and interface states. CX Codex ★ 776repo stars security_auditor security_auditor — a subagent for Claude Code runs defensive application security reviews. Inspects source code, configurations, and dependency scanner logs to report exploitable flaws. CX Codex ★ 776repo stars

Project sponsors

Companies that keep this analytics open