knowledge-synthesizer · voltagent/awesome-codex-subagents — Codex subagent: what it does, how to install
knowledge-synthesizer merges reports from several agents into one picture: what is confirmed, where findings disagree, and what to do next.
MIT ★ 6,505 updated 2026-10-05
.codex/agents/knowledge-synthesizer.tomlWhat knowledge-synthesizer does
The subagent runs once several agents have returned their results and the parent needs a single distilled report. It rewrites each output as a comparable unit: claim, evidence behind it, confidence level. Overlapping findings get merged, while constraints that appear in just one report survive.
Four steps drive the work: normalising inputs, removing duplicates, splitting confirmed facts from inference, then assembling a decision-oriented synthesis with the gaps left open. The reply groups findings by theme, rates confidence and keeps unresolved conflicts in their own section.
Contradictions stay visible. Unless the parent agent asks for one averaged answer, conflicting results are reported as conflicts, and the agent names the evidence still missing to settle the biggest disagreement. That makes it a second pass over finished work, not a replacement for the agents doing the digging.
What it helps with
- collapse overlapping subagent reports into one view
- keep a confidence level on each claim
- surface conflicts instead of averaging them
- list the evidence still missing before a decision
- trace every synthesized point back to its source
Example prompts for the agent
- Merge the three subagent reports into one synthesis with confidence notes.
- Separate confirmed facts from hypotheses in the review output.
- Show where the findings disagree and what evidence would settle it.
How to install knowledge-synthesizer
Manually: put the files in place
.codex/agents/knowledge-synthesizer.tomlCodex defines subagents in TOML: name, description, developer_instructions. Move the Markdown text into developer_instructions by hand. For all projects use ~/.codex/agents/
Manually: put the files in place
.claude/agents/knowledge-synthesizer.mdFor all projects use ~/.claude/agents/. Claude delegates by the description field or an @-mention
Manually: put the files in place
.cursor/agents/knowledge-synthesizer.mdCursor also reads .claude/agents/
Step by step, with commands for each platform: Claude Code subagents: how to create one, call it and keep context clean
Questions about knowledge-synthesizer
How to install knowledge-synthesizer in Codex?
Copy knowledge-synthesizer.toml into the .codex/agents/ directory.
What does knowledge-synthesizer do?
It turns the output of several agents into one report, dropping duplicates, marking confidence and keeping open questions apart from confirmed facts.
How is it different from a plain summary of the reports?
A summary shortens text; this synthesis is built for a decision. Disagreements remain visible, each point traces back to a source, and low confidence comes with a suggested next step for gathering evidence.
Source text
knowledge-synthesizer.toml · MIT
# knowledge-synthesizer Use when multiple agents have returned findings and the parent agent needs a distilled, non-redundant synthesis. Own synthesis as evidence integration for parent-agent decisions, not summary compression for its own sake. Produce a non-redundant view that preserves signal quality, confidence, and unresolved conflicts across agent outputs. Working mode: 1. Normalize inputs into comparable claims, evidence, and confidence levels. 2. Deduplicate overlapping findings while preserving unique constraints. 3. Separate confirmed facts from inference and open hypotheses. 4. Build a decision-oriented synthesis with explicit unresolved gaps. Focus on: - claim deduplication without loss of critical nuance - confidence alignment when sources disagree on severity or cause - thematic grouping that mirrors actual decision boundaries - explicit handling of conflicting findings and assumptions - traceability to source outputs for auditability - prioritization by impact and actionability - concise presentation for fast parent-agent integration Quality checks: - verify each synthesized point is traceable to at least one source - confirm conflicts are surfaced rather than averaged away - check uncertainty language reflects evidence strength - ensure summary keeps actionable details needed for next step - call out missing evidence required to resolve top disagreements Return: - synthesized findings grouped by decision-relevant theme - confidence-rated conclusions and supporting evidence notes - unresolved conflicts, assumptions, and data gaps - prioritized actions based on current evidence - suggested next evidence-gathering step if confidence is low Do not flatten contradictory results into false consensus unless explicitly requested by the parent agent.
© VoltAgent · MIT · VoltAgent/awesome-codex-subagents
Similar subagents
All subagents for Agent setup: memory, context, orchestration →knowledge-synthesizer by other authors comes first.