# EM-0039 execution plan — mission and four-case narrative

- Object ID: `em:documentation:sha256:c9617f8c7a4767c78e6e23a13c69134a766a80a733f46445858879a0025097cb`
- Kind: `documentation`
- Repository path: [`docs/execution-plans/EM-0039.md`](https://github.com/yoheinakajima/epistemedia/blob/f92846570180dfa4511263f8ba98ecd18f7772c9/docs/execution-plans/EM-0039.md)
- Content digest: `0fdbd063db1cf844fe053addb4bedd4d62e6f9044c2708b5ee7ffd3e0b6b0b22`

**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-0039 execution plan — mission and four-case narrative

## Objective

Publish the owner-approved mission as a versioned, inspectable project statement and use it to make the four-case library explain why Epistemedia matters without changing any accepted evidence meaning.

## Authority and invariants

- Task: `EM-0039`
- Accepted base: `97c43a429026000822bb938c9a05b039ab9b592e`
- Accepted dossiers, manifests, and independent review receipts for Cases 001–004 remain byte-for-byte unchanged.
- The failure-mode map is editorial framing, not a new evidence status or universal taxonomy.
- The public agent kit is live only after deployment read-back. Hosted API, hosted MCP submission, and the EM-0038 queue remain unavailable.
- Synthetic tests may validate structure and mechanics. Only retained answers from an actual reader can close the human-comprehension gate.

## Implementation sequence

1. Add one validated, versioned mission source under `catalog/` with an explicit non-policy and non-evidence boundary.
2. Compile deterministic `/about/` HTML, Markdown, and JSON surfaces plus a public reader-check instrument.
3. Keep the featured dossier first, then add a compact mission bridge, the four failure modes, a three-step reading method, and participation paths to the homepage.
4. Recast the How We Know index as four unit tests for how knowledge can fail while preserving every accepted dossier link and policy view.
5. Split reader navigation from project utilities and preserve no-JavaScript keyboard access at desktop and mobile widths.
6. Add fail-closed current/live/future, parity, discovery, accessibility, and accepted-byte regression tests.
7. Run focused checks, full deterministic validation, exact-head independent review, protected merge, Pages deployment, and live release read-back.
8. Present the deployed reader-check to a real human and retain the answers separately before claiming comprehension evidence.

## Evaluation record

- Registration review: PR #65 independently passed and merged as `97c43a429026000822bb938c9a05b039ab9b592e`; resulting-main Validate run `33265045686`, job `99133521778`, succeeded.
- Author implementation: complete. The mission source compiles to `/about/` HTML, Markdown, and JSON; the homepage and library expose four editorial failure modes; primary and utility navigation are separated; and local REST, MCP, CLI, discovery, status, sitemap, and `llms.txt` expose the same mission identity.
- Automated validation: 145 tests passed with a clean public disclosure audit, deterministic comparison build, unchanged accepted source state, and byte-preserved Case 001–004 dossiers and review receipts.
- Responsive review: desktop `1440×900` and mobile `390×844` checks found one H1, zero scripts, no horizontal overflow, visible primary/utility hierarchy, four failure-mode cards, and a 3 px focus outline. Native `<details>/<summary>` click activation passed. Both browser keyboard drivers focused the native summary but did not dispatch Enter activation; retain this as a bounded driver limitation for post-deployment human read-back rather than report it as a pass.
- Independent exact-head review: pending.
- Protected merge and live deployment: pending.
- Human reader result: pending; must not be inferred from automated checks.