agentleFS
Google CloudKnown publisherGoogleCloudPlatform/LifeSciences/applications/open-agentic-science-engine/skills/hypothesis-entry/SKILL.md

hypothesis-entry

Skill53 starsChanged 55 days agoApache-2.0View on GitHub

Hypothesis entry: how hypotheses enter an OASE science program. Four strategies exist, each producing a Layer 0 artifact under raw/hypotheses/ with its own schema and threshold set. Use this skill when planning or executing hypothesis entry — deciding which strategy to use, running the adopt command, or interpreting the assessment output. Hypothesis entry feeds into Stage 0 bounded portfolio triage (see science-program-lead template §5a). Do not use this skill for downstream analysis of hypotheses (stage 1+), for reading co-scientist tournament details (use tournament-corpus), or for running a hypex tournament (use the hypex work-order path).

We split this skill into the rules it gives an agent

1 rule. About 0 are common, 0 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Every strategy's analyze emits the shared oase.hypothesis-assessment.v1 core:

    line 96

≈ “Common” is an estimate: at least 1 in 5 of the 308,232 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

1. The four strategies

StrategyCommandLayer 0 schemaAvailability
Sponsoroase hypothesis adopt --origin sponsoroase.hypothesis-set.v1Always available
Charteroase hypothesis adopt --origin charteroase.hypothesis-set.v1Always available
Co-Scientistoase coscientist ingestoase.coscientist.v1Requires a Co-Scientist export file
Hypexhypex-supervisor work order, then oase hypex ingestoase.hypex.v1Requires OASE-provisioned Hypex tools, templates, and supervisor lease

Each strategy writes its own vendor-native Layer 0 artifact. The shared contract is oase.hypothesis-assessment.v1, emitted by each strategy's analyze command. There is no single score comparable across strategies. score is an object carrying {value, basis} or null; cross-strategy comparison is a judgement the program lead makes in prose.

2. When to prefer each strategy

If the starting material is...UseWhy
A sponsor-supplied list of hypothesesoase hypothesis adopt --origin sponsorHonest provenance; attestation recorded
Hypotheses authored in the program charteroase hypothesis adopt --origin charterSame mechanism, charter-specific infix
A Co-Scientist tournament exportoase coscientist ingestExisting validated pipeline with ELO ranking
A need for adversarial hypothesis explorationhypex-supervisor work orderMulti-epoch tournament with review panel, proximity, and evolution
A published hypothesis set (paper, prior program)oase hypothesis adopt --origin publication or --origin prior-programRequires --cite for the source reference

3. Tool invocations

Adoption (sponsor, charter, prior-program, publication)

# Phase 1: adopt the hypothesis set
oase hypothesis adopt hypotheses.json \
  --origin sponsor \
  --attest "Provided by Dr. Smith on 2026-09-01 as starting material for the CDK4 program" \
  [--cite DOI:10.1234/example]

# Phase 2: assess the adopted set
oase hypothesis analyze raw/hypotheses/<slug>.adopted.json

Co-Scientist (unchanged)

oase coscientist ingest export.json [--top N]
oase coscientist analyze raw/hypotheses/cs-<session>.tournament.json

4. Mandatory relays

Relay codeKindFires whenObligation
hypothesis.adopted_not_generatedQualifierEvery adoptionQuote the attestation verbatim in any finding that rests on this artifact. Do not describe the set as OASE-derived.
hypothesis.unranked_setQualifierEvery adopted set analysisArray position is input order, not preference. Do not present as a leaderboard.

5. The no-silent-substitution rule

If the sponsor named a strategy and oase doctor reports it unavailable, the lead reports blocked. It does NOT silently substitute another.

A substituted strategy produces a differently-shaped artifact with a different score basis. A silent substitution is a provenance error that surfaces much later, when a finding cites an artifact whose shape does not match what the sponsor expected.

Check availability:

oase doctor --json | jq '.checks[] | select(.name | startswith("hypothesis strategy"))'

6. Threshold sets

StrategyThreshold setQuality thresholds?
Adopted (sponsor, charter, prior-program, publication)hypothesis-set@1.0No — structural only (min_candidates)
Co-Scientistcoscientist@1.1Yes — ELO gap, win rate, contradiction count
Hypexhypex@1.0Yes — but three values ship UNRESOLVED

7. Assessment output

Every strategy's analyze emits the shared oase.hypothesis-assessment.v1 core:

{
  "schema": "oase.hypothesis-assessment.v1",
  "strategy": "adopted | coscientist | hypex",
  "source_artifact": "raw/hypotheses/...",
  "source_sha256": "...",
  "candidates": [
    {
      "candidate_id": "...",
      "statement": "...",
      "rank": null,
      "score": null,
      "origin": "adopted | generated"
    }
  ]
}

For adopted sets: rank is always null, score is always null. For ranked strategies: score is {value: <number>, basis: "<set>"} — never a bare number.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.