dbm-postauthor-concordance

name	dbm-postauthor-concordance
description	Post-authoring evidence bundle and review for pipeline-produced DBM sections — build review substrate, prepare candidate findings for human readiness judgment.
compatibility	Chirality TASK; dispatched by DBM_PUBLISHER Gate 6
allowed-tools	python3 tools/review/scan_section_coverage.py:*, python3 tools/review/extract_claims.py:*, python3 tools/review/scan_tbd_markers.py:*, python3 tools/review/check_body_thinness.py:*
metadata	{"chirality-skill-version":"1","chirality-task-profile":"NONE"}

SKILL - dbm-postauthor-concordance

Purpose

Post-authoring review for pipeline-produced DBM sections. Build a structured evidence bundle using deterministic substrate tools, then use agent judgment to read governed truth alongside the authored DBM and prepare candidate findings for human readiness judgment.

This skill replaces the pre-authoring concordance model (Gates 3-4 concordance freeze). Concordance is now extracted and validated after authoring rather than constraining authoring. The evidence bundle is not a quality gate — findings are candidates for human disposition, not automatic blockers.

This skill is a reviewer only. It does not modify section outputs, the knowledge base, publication planning artifacts, or governed pointers.

Suitable agent shells

• TASK (generic shell mode, no profile)

Typical dispatcher: DBM_PUBLISHER during Gate 6 after section synthesis and package assembly.

Inputs

Required

• ASSEMBLED_DBM_PATH — path to the assembled Rewritten_DBM.md (or sections root)
• REVIEW_OUTPUT_DIR — output directory for the evidence bundle
• DOMAIN_ROOT — absolute path to the DOMAIN root
• PUBLICATION_SCHEMA_PATH — approved Publication_Schema.md
• SECTION_MAP_PATH — approved Section_Map.csv
• PUBLICATION_RULES_PATH — approved Publication_Rules.md

Optional

• SUPERSESSION_MAP_PATH — Supersession_Map.csv from active SCA
• SECTION_CONTEXT_ROOT — directory containing SEC-##_Context.md packets
• SECTIONS_ROOT — individual section output folders (for section-level QA artifact access)

Runtime overrides

Key	Meaning	Default	Allowed values
ASSEMBLED_DBM_PATH	Assembled DBM to review	Required	Absolute path to Rewritten_DBM.md
REVIEW_OUTPUT_DIR	Evidence bundle output root	Required	Path under _Publication/DBM/package/<run-id>/review/
DOMAIN_ROOT	DOMAIN decomposition root	Required	Absolute path to domain directory
PUBLICATION_SCHEMA_PATH	Publication schema	Required	Markdown path under _Planning/
SECTION_MAP_PATH	Section map	Required	CSV path under _Planning/
PUBLICATION_RULES_PATH	Publication rules	Required	Markdown path under _Planning/
SUPERSESSION_MAP_PATH	Supersession map	unset	CSV path under active SCA
SECTION_CONTEXT_ROOT	Section context packets	unset	Directory path
SECTIONS_ROOT	Section output folders	unset	Path under _Publication/DBM/sections/

Two-level status model

EvidenceBundleStatus (process validity)

Status	Meaning
COMPLETE	All tools ran, all governed inputs available, all substrate artifacts written
PARTIAL	One or more governed inputs unavailable; substrate is incomplete
TOOL_ERROR	A tool failed; substrate is unreliable

Expected to be COMPLETE in valid pipeline runs, but the skill must still handle PARTIAL and TOOL_ERROR — files can be missing, malformed, stale, or unreadable.

ReviewStatus (agent candidate-finding result)

Status	Meaning
NO_FINDINGS	Agent review produced no candidate findings
FINDINGS_FOR_DISPOSITION	Agent prepared candidate findings for human disposition
REVIEW_INCOMPLETE	Agent could not complete review

Finding taxonomy (controlled enum)

Type	Meaning
INCORRECT	Authored content contradicts governed KTY/SCA/supersession truth
UNSUPPORTED	Authored content makes a claim not warranted by the knowledge base
MISSING	Governed material exists but is absent from the authored content
FLATTENED	Authored content converts TBD/conflict/assumption into a firm fact
OUTDATED	Authored content uses superseded source DBM content
INCOMPLETE	Authored content has the right topic but lacks required engineering detail

Finding severity (controlled enum)

Severity	Meaning
HIGH	Significant technical or governance concern
MEDIUM	Notable gap or inconsistency
LOW	Minor quality observation
ADVISORY	Noted for record; no action expected

Method

Tool-first: build review substrate

1. Validate inputs and write boundary. Confirm all required paths exist. Create output directory if absent. If any required input is missing or unreadable, set EvidenceBundleStatus = PARTIAL and record the gap.

2. Run scan_section_coverage.py. Pass --section-map (required for pipeline review — Section_Map.csv is the run-specific authority after Gate 4). Write output to {REVIEW_OUTPUT_DIR}/Section_Coverage.csv.

3. Run extract_claims.py. Extract engineering values, parameters, configuration statements from assembled DBM. Write output to {REVIEW_OUTPUT_DIR}/Draft_Claims.csv.

4. Run scan_tbd_markers.py. Scan assembled DBM for TBD/TBC/ASSUMPTION markers with KB cross-reference. Write output to {REVIEW_OUTPUT_DIR}/TBD_Inventory.csv.

5. Run check_body_thinness.py. Pass --section-map and --schema. Write output to {REVIEW_OUTPUT_DIR}/Body_Thinness.csv.

Agent judgment: review governed truth against authored content

6. Read authored sections against mapped KA artifacts. Use Section_Map.csv to identify PRIMARY-role KA artifacts per section. For each section, judge materiality, accuracy, and completeness. This is the core engineering judgment step.

7. Check supersession compliance. When SUPERSESSION_MAP_PATH is provided, verify authored content uses current values.

8. Classify candidate findings. Every finding must set Origin = AGENT_CHECK, cite governed-truth evidence, and trace to substrate where applicable.

9. Emit evidence bundle and disposition artifact. Write all outputs to REVIEW_OUTPUT_DIR. Write Publication_Review_Disposition.csv with HumanDisposition = TBD for all rows.

Outputs

All outputs are written to REVIEW_OUTPUT_DIR. 7 files total:

File	Purpose
Evidence_Bundle_Summary.md	EvidenceBundleStatus, ReviewStatus, tool run results, input provenance
Section_Coverage.csv	Structural coverage
Draft_Claims.csv	Values, parameters, terms located in authored text
Body_Thinness.csv	Section body underdevelopment signals
TBD_Inventory.csv	TBD/TBC/ASSUMPTION markers with KB cross-reference
Candidate_Findings.csv	Agent-prepared findings for human disposition
Publication_Review_Disposition.csv	Disposition artifact for human decision recording

Candidate_Findings.csv schema

Same as dbm-draft-review:

FindingID, FindingType, Severity, Origin, DraftLocation, DraftLineNumber, DraftText, GovernedTruthRef, GovernedTruthValue, SectionID, KTYRef, KARef, SupersessionRef, Explanation, EvidenceSource

Publication_Review_Disposition.csv schema

FindingID, FindingType, Severity, SectionID, Explanation, EvidenceSource, ProposedDisposition, HumanDisposition, HumanNotes, DispositionDate

• FindingID: matches Candidate_Findings.csv FindingID
• ProposedDisposition: agent-recommended disposition
• HumanDisposition: TBD until human rules. Allowed values: ACCEPT_AS_IS, REVISE, WAIVE_WITH_RATIONALE, DEFER, NOT_APPLICABLE
• HumanNotes: empty until human fills in
• DispositionDate: empty until human fills in

DBM_PUBLISHER does not update acceptance pointers until all HumanDisposition values are recorded (no TBD remaining).

Evidence_Bundle_Summary.md structure

Same required H2 headings as dbm-draft-review.

Non-negotiable constraints

• The skill does NOT modify section outputs, knowledge base, publication artifacts, or governed pointers.
• Write boundary is exactly the 7 output files under REVIEW_OUTPUT_DIR.
• No finding is treated as accepted truth. Findings are candidates for human disposition.
• Origin = AGENT_CHECK on every finding row.
• No automatic readiness verdicts, PASS/FAIL judgments, or blocking decisions.
• Publication_Review_Disposition.csv is emitted with HumanDisposition = TBD — the skill does not pre-fill human decisions.

QA expectations

• All 7 required outputs exist under REVIEW_OUTPUT_DIR.
• All writes stayed inside REVIEW_OUTPUT_DIR.
• Candidate_Findings.csv uses only controlled enum values for FindingType and Severity.
• Every finding has Origin = AGENT_CHECK, non-empty Explanation and EvidenceSource.
• Publication_Review_Disposition.csv has one row per finding, with HumanDisposition = TBD.
• Evidence_Bundle_Summary.md contains both EvidenceBundleStatus and ReviewStatus.