zrwangems/dsh-zrwang-plugin / ars/skills/deep-research
deep-research
Flagged学术研究前置:把模糊想法收敛成可研究的问题,并做文献综述、事实核查、系统综述(PRISMA)与三向文献扫描。Triggers: research, deep research, literature review, systematic review, meta-analysis, PRISMA, fact-check, guide my research | 研究, 文献综述, 系统性文献回顾, 事实核查, 引导我研究, 帮我确定研究问题。输出 RQ Brief、方法论蓝图、带核验的注释书目。
_shared/contracts/passport/claim_intent_manifest.schema.json
View on GitHub{
"$schema": "https://json-schema.org/draft/2020-12/schema",
"$id": "https://github.com/Imbad0202/academic-research-skills/shared/contracts/passport/claim_intent_manifest.schema.json",
"title": "Material Passport Claim Intent Manifest Entry",
"description": "One entry in Material Passport Schema 9 claim_intent_manifests[] aggregate. Emitted by synthesis_agent / draft_writer_agent / report_compiler_agent after paper-visible context loads but before prose generation, per the 'Claim Intent Manifest Emission (v3.8)' sibling section added to each writing-stage agent. Acts as the pre-commitment baseline for the manifest set-diff in claim_ref_alignment_audit_agent §4 step 5 (intended vs emitted vs supported claims). Cross-field invariants M-INV-1..M-INV-4 are lint-enforced in scripts/check_claim_audit_consistency.py. See docs/design/2026-05-15-issue-103-claim-alignment-audit-spec.md §3.2.",
"type": "object",
"additionalProperties": false,
"required": [
"manifest_version",
"manifest_id",
"emitted_by",
"emitted_at",
"claims",
"manifest_negative_constraints"
],
"properties": {
"manifest_version": {
"const": "1.0",
"description": "Schema-internal version. v3.8.0 release ships 1.0. Future drafts bump this constant to drive an explicit migration."
},
"manifest_id": {
"type": "string",
"pattern": "^M-[0-9]{4}-[0-9]{2}-[0-9]{2}T[0-9]{2}:[0-9]{2}:[0-9]{2}Z-[0-9a-f]{4}$",
"description": "Discriminator scoping all claim_id values inside this manifest. Format: M-<ISO-8601-Z>-<4-hex>. Required because a single passport may contain multiple claim_intent_manifests[] entries (e.g., one from synthesis_agent and one from draft_writer_agent on the same run); bare C-001 alone would collide. The pair (manifest_id, claim_id) is the joinable key — see claim_audit_result.scoped_manifest_id + INV-15 / D-INV-2 cross-array integrity. M-INV-4 enforces uniqueness across all claim_intent_manifests[] entries in the same passport."
},
"emitted_by": {
"enum": ["synthesis_agent", "draft_writer_agent", "report_compiler_agent"],
"description": "The writing-stage agent that emitted this manifest. Bibliography / audit / orchestrator agents do NOT emit manifests — only the three prose-producing agents pre-commit to claim intent."
},
"emitted_at": { "type": "string", "format": "date-time" },
"session_id": {
"type": "string",
"description": "Optional opaque identifier for the dispatch run. Useful for cross-referencing with orchestrator logs; not part of the joinable key."
},
"claims": {
"type": "array",
"minItems": 1,
"description": "MUST list at least one substantive claim per R-CIM-A — an empty manifest is a degenerate pre-commitment that produces no drift signal (set-diff has nothing in Intended set) and silently defeats D6.",
"items": {
"type": "object",
"additionalProperties": false,
"required": ["claim_id", "claim_text", "intended_evidence_kind", "planned_refs"],
"properties": {
"claim_id": {
"type": "string",
"pattern": "^C-[0-9]{3,}$",
"description": "Per-manifest claim identifier. Unique within this manifest (M-INV-1); cross-manifest collisions are PERMITTED and disambiguated via the scoping pair (manifest_id, claim_id)."
},
"claim_text": { "type": "string", "minLength": 1, "maxLength": 2000 },
"intended_evidence_kind": {
"enum": ["empirical", "theoretical", "definitional", "normative"],
"description": "What kind of evidence the writer intends to back this claim with. Used by the audit agent's manifest cross-reference to discriminate drift from refinement."
},
"planned_refs": {
"type": "array",
"items": { "type": "string" },
"minItems": 0,
"description": "Ref slugs the writer intends to cite. Empty array MUST be permitted — definitional/normative claims may legitimately have no planned refs at manifest time."
},
"planned_experiment_ids": {
"type": "array",
"items": { "type": "string" },
"minItems": 1,
"uniqueItems": true,
"description": "Passport-local experiment_provenance[].experiment_id values the writer intends to support this claim with (#260, D1). OPTIONAL-ABSENT — present only on experiment-backed claims; literature-only / definitional / theoretical / normative claims omit it entirely (do NOT carry a meaningless empty array). minItems 1 because an empty array is meaningless — omit the field instead. EP-INV-2 enforces every value resolves to exactly one experiment_provenance[] entry (doubles as the rename + forward-reference dangling-pointer guard). EP-INV-3 enforces presence ⟹ this claim's intended_evidence_kind == 'empirical' (mixed literature+experiment allowed: a claim MAY carry BOTH planned_refs and planned_experiment_ids). A mixed-evidence claim is audited by both paths and combined worst-verdict-wins (D4)."
},
"negative_constraints": {
"type": "array",
"items": {
"type": "object",
"additionalProperties": false,
"required": ["constraint_id", "rule"],
"properties": {
"constraint_id": {
"type": "string",
"pattern": "^NC-C[0-9]{3,}-[0-9]+$",
"description": "Claim-level negative constraint. Format NC-C{n}-{m} where {n} matches the parent claim_id digits (M-INV-2 + INV-17). NO inner hyphen between 'C' and '{n}' — the joinable key is the digit sequence after stripping the leading 'C'."
},
"rule": { "type": "string", "minLength": 1, "maxLength": 500 }
}
}
}
}
}
},
"manifest_negative_constraints": {
"type": "array",
"description": "Globally-scoped constraints. Apply to all claims in this manifest regardless of whether the claim entry redeclares them (M-INV-3 — claim-level can ADD, never DROP global).",
"items": {
"type": "object",
"additionalProperties": false,
"required": ["constraint_id", "rule"],
"properties": {
"constraint_id": {
"type": "string",
"pattern": "^MNC-[0-9]+$",
"description": "Manifest-level negative constraint. Format MNC-{m}. Globally applied; cannot be overridden by claim-level NC."
},
"rule": { "type": "string", "minLength": 1, "maxLength": 500 }
}
}
}
}
}
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| add-product-e2e-evalpaperclipai/paperclipAdd or extend a Paperclip full-stack runner E2E workflow, fixture, matcher, or report evidence path for local or Daytona execution. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| add-runner-evalpaperclipai/paperclipAdd or extend a Paperclip Runner protocol evaluation definition, roster, assertion, or report fixture with provenance and narrow validation. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| diagnose-why-work-stoppedpaperclipai/paperclipCompatibility entry point for failed, stalled, looping, or over-recovered Paperclip work. Use fix-failed-paperclip-run for the full diagnosis, regression, repair, real verification, and green-PR workflow; honor an explicitly requested diagnosis-only scope. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| fix-failed-paperclip-runpaperclipai/paperclipDiagnose and fix failed, stuck, slow, or incorrectly blocked Paperclip tasks and runs. Use with a task/run URL or failure report to trace the root cause, prove a regression red then green, check second-order effects, verify the real workflow, and deliver a PR with passing checks. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |