web-consensus
Use when a question has to be settled against the live web rather than from memory - researching a topic, proving or disproving a claim, checking whether something is still true, a sourced second opinion on a factual question - or when the user says "research", "prove it", "with sources", "check if this is true", "still accurate?".
- License 许可
- Apache-2.0
- 最近更新
- 2026-11-14
- Stars
- ★ 0
- Forks
- 0
01 前言 Front Matter
- name
- web-consensus
- description
- Use when a question has to be settled against the live web rather than from memory - researching a topic, proving or disproving a claim, checking whether something is still true, a sourced second opinion on a factual question - or when the user says "research", "prove it", "with sources", "check if this is true", "still accurate?".
02 资源文件 1
- 其他.codex-harness6 B
03 SKILL.md 正文
Web consensus
Cheap researchers gather evidence in parallel, a reviewer from another model family attacks it, the disputed points are researched again from scratch, a script checks every link and quote, and a judge on the strongest model writes the verdict from sources only.
Who does what
| Step | Subagent role | Runs on |
|---|---|---|
| Evidence, cold verification | explorer, one per angle or question, in parallel |
bulk tier: fast and cheap |
| Challenge | reviewer |
review tier: another model family |
| Link and quote check | harness check-sources |
no model |
| Verdict | judge |
deep tier: the strongest model |
This skill asks for subagents: every role in the table runs through spawn_agent with that agent_type, even when you wait for its result before the next step, because the point is a different model and a fresh context, not speed. Each subagent gets its own brief and nothing about the others, so no role can quietly agree with another, and never a fork of this conversation. You only route: write the evidence files, run the script, pass material between roles. Run the whole flow even when the answer looks obvious or one page settles it: "it's a simple fact", "the docs already say it" and "one search is enough" are the answers from memory this skill replaces. If spawn_agent is not available in this session, stop and say so: a consensus from one model playing every role is not this skill's output, so do not play the roles yourself.
Flow
-
Evidence. One
explorerper angle. Each returns claim / URL / exact quote triples and nothing else. Write them toevidence-<angle>.jsonlin the workspace, one JSON object per line:{"claim": "...", "url": "...", "quote": "..."}.direct: the question as asked.disconfirming: the strongest evidence that the expected answer is wrong: corrections, retractions, errata, dissenting primary sources.primary-source: official docs, specifications, source code, changelogs and papers only.academic, when the question has a research literature: papers, reproductions, negative results.
Vary the wording of the searches between angles. The same index returns very different results for differently worded questions.
-
Check.
harness check-sources "evidence-*.jsonl"(the command expands the pattern itself, so the same line works in PowerShell). It fetches each URL once and reports per claim:ok(the page contains the quote),quote-missing,dead,blocked(401/403/429: usually bot-blocking, not absence),unreadable(PDF or binary: quote unverified). It needs the network: in a sandbox, ask for escalation on the first run, not after a refusal. When the workspace is read-only, pipe the lines in and pass-as the file (printf '%s\n' "$line1" "$line2" | harness check-sources -, in PowerShell$lines | harness check-sources -), and hand the evidence to the later roles inline in their briefs. One run covers every file. When it runs, the final reply quotes its summary line (N claims: ...) exactly as printed. If it still cannot run, nothing is checked: say so, the verdict's first line isUNVERIFIED:and no claim is agreed. -
Challenge. A
reviewerreads the evidence files and the check report, and is told not to search. It rules on every claim with one line each,OK,DOUBTorMISMATCH, then lists the questions that would settle each dispute. -
Cold verification. For each of those questions, a fresh
explorergets only the question: no evidence and no challenge. Write its triples toevidence-cold.jsonland run the check on it. When the challenge lists no disputes, skip this step and say so in the report; there is then only the first check report. -
Verdict. A
judgegets the question, every evidence file, every check report (one, or two after cold verification) and the challenge. It answers in a short paragraph, then lists agreed, disputed and unsourced claims with URL and quote, its confidence, and what new evidence would change the answer. Relay the verdict; do not rewrite it, and never add that sources are independent.
Rules
- A claim can only be agreed when
check-sourcessaysok, orunreadablewith the quote confirmed by a second source. - The searcher never dictates the answer alone. Evidence files from the same search are not independent sources; say how many distinct sites agree.
- Prefer primary sources over secondary ones, and recent over old when the subject moves.
- When results look stale, ask the user to start the session with live search (
codex -c web_search="live"). - If web search is not available in this session, say so and stop. Do not answer from memory.
Common mistakes
| Mistake | Instead |
|---|---|
| Answering from one search because the question looks easy | Run every step; the user asked for sources checked by roles that did not write them. |
| Forking this conversation into a researcher | A fresh brief; a fork re-reads everything and inherits your bias. |
| Judging in the main session | The judge role exists so the verdict runs on the strongest model whatever tier this session uses. |
Treating blocked as dead |
Say it was blocked; find a second source or quote it as unverified. |
Paraphrasing in the quote field |
The check needs the exact words on the page. |
04 同仓库更多 Skills
skill-validate
Use when a SKILL.md is about to ship or was just edited, when a skill is reviewed or audited, when a skill seems ignored, skipped or misread by agents, or when the user says "validate this skill", "lint the skill", "check my skill", "is this skill any good".
regression-tests
Use when fixing a bug, changing behaviour or adding a feature in code that has tests, or when the user asks whether a change is covered by tests, wants regression tests, or wants coverage of the changed lines.
redacao
Use when writing, rewriting or editing a technical article or blog post in Portuguese, or when the user says "escreve um artigo", "reescreve o post", "redação", "pauta", "revisa o texto", "tira a cara de IA".
improve
Use when the user asks to improve the harness, its rules, Guardian policy, skills or knowledge from experience on this machine, or says "remember this for next time", "make this a rule", "teach the harness".
hld-trace
Use when the user asks whether the code follows the high-level design (HLD), which requirements have tests, which are untested, or wants tests linked to requirement or work item IDs.
grilling
Use when the user wants a plan, decision, design or idea stress-tested before committing to it, or says "grill me", "grill this", "stress-test this", "poke holes in this", "what am I missing". Do not fire on your own.