# EM-0023 execution plan

- Object ID: `em:documentation:sha256:038b4f8106811df85bb9c74a5153006eaadbf0d78f5d83ac46b3510b5bf4dd34`
- Kind: `documentation`
- Repository path: [`docs/execution-plans/EM-0023.md`](https://github.com/yoheinakajima/epistemedia/blob/f92846570180dfa4511263f8ba98ecd18f7772c9/docs/execution-plans/EM-0023.md)
- Content digest: `b9546743fba0579126045dec1f2ef53741a1c72bf6c7b22649b0ae6d2a62f5c7`

**Also filed under:** [Disclosure and Public Projection](https://epistemedia.org/topics/disclosure/), [Epistemedia](https://epistemedia.org/topics/epistemedia/), [Epistemic Mesh Protocol](https://epistemedia.org/topics/epistemic-mesh/), [Sovereign Realm Federation](https://epistemedia.org/topics/federation/), [Autonomous Governance](https://epistemedia.org/topics/governance/), [Knowledge Objects](https://epistemedia.org/topics/knowledge-objects/), [Human and Agent Interfaces](https://epistemedia.org/topics/public-interfaces/), [Releases and Reproducibility](https://epistemedia.org/topics/releases/), [Research Program](https://epistemedia.org/topics/research-program/), [Security and Adversarial Robustness](https://epistemedia.org/topics/security/)

## Source content

# EM-0023 execution plan

Status: candidate complete on `codex/em-0023-living-evidence-report`; exact-head independent
review pending.

## Objective

Turn Case 001 from a flat rendered report into a living forensic-editorial composition. Preserve
the accepted dossier while giving the finding, Stranger Test, practical meaning, complete evidence
ledger, argument sequence, unresolved lineage, terminology, and launch-ready scoreboard distinct
visual and semantic roles.

## Change surface

1. Derive complete assertion and participant-data-root ledger entries from the same accepted
   relations, lineages, assertions, spans, editions, and source works that produce the scoreboard.
2. Make all three scoreboard cells link to those complete lists and define the Stranger Test in one
   plain sentence.
3. Compile a visible five-term lexicon and a qualified, policy-relative practical reading with
   explicit relation/span closure for each promoted view.
4. Recompose the report into lede, evidence scoreboard, annotated relation sequence, root ledger,
   terminology, uncertainty, and source-register movements with a stronger typographic rhythm.
5. Generate deterministic view-specific SVG share cards from the accepted projection identity,
   question, verdict, counts, and canonical route.

## Hard boundaries

- No new case, source, evidence assertion, lineage judgment, count, relation meaning, receipt, or
  research-packet change.
- The practical reading remains visibly qualified and policy-relative; it is not universal
  professional or medical advice.
- Encyclopedia and skeptical remain the only promoted views and continue to compile from one
  byte-identical dossier.
- No workflow, DNS, hosted API/MCP, package, release, or unrelated repository work.

## Verification

- exact ledger-count/key/source derivation and displayed-list closure;
- practical-reading relation/span closure and view divergence;
- deterministic SVG bytes, escaping, identities, canonical routes, and content types;
- static/API/MCP/CLI projection parity;
- no-script semantic hierarchy, native links/details, focus, contrast, and reduced motion;
- desktop and mobile inspection at representative wide and narrow viewports;
- full `make check`, exact-head independent review, protected merge, Pages deployment, and live
  route/content-type/identity read-back.

## Candidate verification

- Targeted Ruff and interface/featured tests passed.
- `make check PYTHON=.venv/bin/python` passed 78 tests, accepted-input validation, public
  disclosure audit with zero findings, independent deterministic rebuild, and unchanged tracked
  source state. Local run receipt:
  `runs/proposals/20260822T231026Z-f63155d87b6c.json`.
- A local Pages-equivalent build produced 569 inventoried files. The encyclopedia and skeptical
  HTML, Markdown, JSON, and SVG card projections compiled from the same accepted dossier.
- Browser inspection at 1440 by 900 and 390 by 844 found one H1, zero scripts, no horizontal
  overflow, 27 displayed evidence-ledger entries, five glossary terms, and seven native details
  controls. The encyclopedia and skeptical first screens expose different verdicts and practical
  readings. The ledger and lexicon retain readable hierarchy at the narrow viewport.
- The generated 1200 by 630 SVG card visibly carries the question, policy view, verdict, all three
  scoreboard counts, dossier identity, and canonical route.
- Accepted research inputs remain byte-identical: dossier SHA-256 `7003413e286e4d310f81441db33f4a467ba2eb3e08f41ddfa3cef5abb34707ca`
  (214,499 bytes), independent receipt SHA-256
  `503d16396b25b1c22d7fc10ac6fb7db2e530e6ce348d63fa8b639db5a5288f0a`
  (65,611 bytes), and feature manifest SHA-256
  `5c96dead036b527793ba5a0de59bf7316efdfeb591470d4a23e5bf979f3b9288`
  (2,205 bytes).

The first independent review bound head `8770de9f5264341217b83e3c6e5afc777f852874`
and required two link corrections: restore a direct first-screen route to the unresolved 2007
lineage while retaining the complete support-root ledger, and keep tally fragments inside the
selected skeptical view. Both were corrected in a forward commit with regression tests. The
post-correction full gate passed again; run receipt:
`runs/proposals/20260822T232023Z-7976b9a7a1c3.json`. Mobile read-back at 390 by 844 confirmed the
direct route lands with “What remains unresolved” visible, and all skeptical tally links are local
fragments.