fix: architecture review corrections — CONTEXT.md trim and eval backtrack verification

- Remove Fleet, Symbolic oversight, and Data classification tiers from
  CONTEXT.md glossary: Fleet is Phase 2 only; the other two duplicate
  always-on governance.md content
- Drop now-stale governance test assertions for the removed terms
- write-skill eval: add section-ordering assertion; fix duplicate prompt
  across two deterministic tests
- write-eval eval: tighten plan-shown assertion to require per-case
  detail (query + type), not just "a list of test cases"

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

.agents/evals/factory/write-eval/eval.yaml

@@ -58,9 +58,9 @@ output_tests:
     name: "Deterministic — test plan presented before file is written"
     type: deterministic
     prompt: "write evals for the diagnose skill"
-    expected_output: "Skill presents proposed test cases for review and requests confirmation before writing any file"
+    expected_output: "Skill presents each proposed test case with its id, type, and query before writing any file, then requests confirmation"
     assertions:
-      - "Response includes a proposed test plan or list of test cases before any file is written"
+      - "Response presents each proposed test case individually — showing at minimum the query and test type — before any file is written"
       - "Response requests confirmation before proceeding to write"
 
   - id: deterministic-merge-conflict-flagged

.agents/evals/factory/write-skill/eval.yaml

@@ -33,10 +33,10 @@ trigger_tests:
 
 output_tests:
   - id: output-has-all-sections
-    name: All 8 body sections present
+    name: All 8 body sections present in order
     type: deterministic
     prompt: "Write a new skill for linting markdown files, category: implement"
-    expected_output: A complete SKILL.md containing all 8 required body sections in order.
+    expected_output: A complete SKILL.md containing all 8 required body sections in the prescribed order.
     assertions:
       - "Output contains '## Role'"
       - "Output contains '## When to use / When not to use'"
@@ -46,11 +46,12 @@ output_tests:
       - "Output contains '## Output format'"
       - "Output contains '## Failure handling'"
       - "Output contains '## Self-check'"
+      - "Sections appear in this order: ## Role, ## When to use / When not to use, ## Required inputs, ## Constraints, ## Process, ## Output format, ## Failure handling, ## Self-check"
 
   - id: output-path-correct
     name: Output path and frontmatter fields correct
     type: deterministic
-    prompt: "Write a new skill for linting markdown files, category: implement"
+    prompt: "Write a new skill for sending Slack notifications on deploy events, category: deploy"
     expected_output: A SKILL.md with correct output path stated and all required frontmatter fields present.
     assertions:
       - "Output contains '.agents/skills/' in the stated output path"

CONTEXT.md

@@ -64,9 +64,6 @@ This repo is built by a junior developer as a homelab tool intended to scale to
 ### Management Application
 A separate product (separate repo) for browsing, editing, and configuring AI development configs through a proper product UI. Git is the persistence layer, invisible to the user. The app is repo-agnostic — it works with any git repo that follows these conventions. This repo is the canonical default content (the official starter). See `docs/VISION.md` for the phased roadmap.
 
-### Fleet
-The set of machines and projects under management in Phase 2 of the Management Application. A fleet member is any machine that has the config deployed and can be reached by the runtime orchestration layer.
-
 ### Skills
 Reusable slash commands for AI coding tools, defined as `SKILL.md` files following the [Agent Skills open standard](https://agentskills.io). Canonical location: `.agents/skills/<skill-name>/SKILL.md` in this repo; deployed to `~/.agents/skills/` on install. Providers that don't read `~/.agents/skills/` natively get a symlink adapter declared in `providers/<name>/provider-manifest.sh` (e.g. Claude Code: `~/.claude/skills/ → ~/.agents/skills/`).
 
@@ -82,21 +79,6 @@ Agent pauses before a consequential action; human approves before execution. Req
 ### HOTL (human-on-the-loop)
 Agent acts; human monitors and can intervene after the fact. Acceptable for low-stakes, bounded, reversible actions where the cost of pausing for approval exceeds the blast radius of an error. The distinction between HITL and HOTL must be explicit and documented — defaulting to HOTL for convenience is not acceptable.
 
-### Symbolic oversight
-Oversight implemented as a gesture (assigning a reviewer) rather than a functional safeguard. A reviewer without the information, time, agency, or intent to evaluate is not oversight — it is the appearance of oversight. The documented failure mode: symbolic oversight passes audits but does not catch errors. Genuine oversight requires: the reviewer has access to what was produced, time to evaluate it meaningfully, authority to reject it, and the intent to do so.
-
-### Data classification tiers
-The four-tier framework governing what data may enter AI context. Apply the tier of the most sensitive element in any dataset or prompt.
-
-| Tier | Examples | AI Rule |
-|---|---|---|
-| **Public** | Publicly available info | No restrictions |
-| **Internal** | Operational data, anonymised logs | Enterprise AI tools only; not consumer/free-tier |
-| **Confidential** | Source code, architecture, personal data, IP | Enterprise AI + contractual data-not-trained guarantee |
-| **Restricted** | GDPR Article 9 health/biometrics, credentials, regulated financial data | Never enters AI context — hard architectural prohibition |
-
-Defined in full in `docs/ai-constitution.md` Section 3. Agent-actionable rules in `core/instructions/governance.md`.
-
 ### Sycophancy
 The failure mode where RLHF-trained models prioritise approval over accuracy. Treated as a first-class reliability risk: models change correct answers to wrong ones under user pressure in a majority of observed cases, then persist in the wrong answer. Designing against sycophancy is an explicit obligation, not a quality-of-life concern. Countermeasures: explicit pushback resistance instructions, prompting for dissent, cross-validating against independent sources. Never interpret AI agreement as AI accuracy.
 

tests/test-governance-layer.sh

@@ -111,14 +111,6 @@ contains "HOTL" "$CONTEXT" \
   && pass "CONTEXT.md: HOTL term defined" \
   || fail "CONTEXT.md: HOTL definition missing"
 
-contains "[Ss]ymbolic oversight" "$CONTEXT" \
-  && pass "CONTEXT.md: Symbolic oversight defined" \
-  || fail "CONTEXT.md: Symbolic oversight definition missing"
-
-contains "[Dd]ata classification" "$CONTEXT" \
-  && pass "CONTEXT.md: Data classification tiers defined" \
-  || fail "CONTEXT.md: Data classification tiers missing"
-
 contains "[Ss]ycophancy" "$CONTEXT" \
   && pass "CONTEXT.md: Sycophancy defined" \
   || fail "CONTEXT.md: Sycophancy definition missing"