Harden Codex parity and installer safety

Author: Cipher-85

.agents/skills/architecture-decision/SKILL.md

@@ -334,7 +334,7 @@ to implement it.]
 5.5. **Engine Specialist Validation** — Before saving, spawn the **primary engine specialist** via Codex subagent delegation to validate the drafted ADR:
    - Read `.codex/docs/technical-preferences.md` `Engine Specialists` section to get the primary specialist
    - If no engine is configured (`[TO BE CONFIGURED]`), skip this step
-   - Spawn `subagent_type: [primary specialist]` with: the ADR's Engine Compatibility section, Decision section, Key Interfaces, and the engine reference docs path. Ask them to:
+   - Spawn `[primary specialist]` via Codex subagent delegation with: the ADR's Engine Compatibility section, Decision section, Key Interfaces, and the engine reference docs path. Ask them to:
      1. Confirm the proposed approach is idiomatic for the pinned engine version
      2. Flag any APIs or patterns that are deprecated or changed post-training-cutoff
      3. Identify engine-specific risks or gotchas not captured in the current ADR draft

.agents/skills/architecture-review/SKILL.md

@@ -321,7 +321,7 @@ Post-Cutoff API Conflicts:
 After completing the engine audit above, spawn the **primary engine specialist** via Codex subagent delegation for a domain-expert second opinion:
 - Read `.codex/docs/technical-preferences.md` `Engine Specialists` section to get the primary specialist
 - If no engine is configured, skip this consultation
-- Spawn `subagent_type: [primary specialist]` with: all ADRs that contain engine-specific decisions or `Post-Cutoff APIs Used` fields, the engine reference docs, and the Phase 5 audit findings. Ask them to:
+- Spawn `[primary specialist]` via Codex subagent delegation with: all ADRs that contain engine-specific decisions or `Post-Cutoff APIs Used` fields, the engine reference docs, and the Phase 5 audit findings. Ask them to:
   1. Confirm or challenge each audit finding — specialists may know of engine nuances not captured in the reference docs
   2. Identify engine-specific anti-patterns in the ADRs that the audit may have missed (e.g., using the wrong Godot node type, Unity component coupling, Unreal subsystem misuse)
   3. Flag ADRs that make assumptions about engine behaviour that differ from the actual pinned version

.agents/skills/handoff/SKILL.md

@@ -252,3 +252,7 @@ Report in 15 lines or fewer:
 - Open blockers or deferred items.
 
 After reporting, stop. Do not start new feature work in the same turn.
+
+## Ported metadata
+
+This is a Codex-native support skill. It has no upstream Claude skill equivalent.

.agents/skills/help/SKILL.md

@@ -43,8 +43,8 @@ Collect these for the output in Step 7 — show them as a footer block:
 
 ```
 ### Also installed (not in workflow)
-- `/skill-name` — [description from SKILL.md frontmatter]
-- `/skill-name` — [description]
+- `$skill-name` — [description from SKILL.md frontmatter]
+- `$skill-name` — [description]
 ```
 
 Only show this block if at least one uncataloged skill exists. Limit to the 10

.agents/skills/hotfix/SKILL.md

@@ -95,9 +95,9 @@ Update the hotfix record with root cause, fix details, and test results.
 
 Use the Codex subagent delegation to request sign-off in parallel:
 
-- `subagent_type: lead-programmer` — Review the fix for correctness and side effects
-- `subagent_type: qa-tester` — Run targeted regression tests on the affected system
-- `subagent_type: producer` — Approve deployment timing and communication plan
+- Delegate to `lead-programmer` — Review the fix for correctness and side effects
+- Delegate to `qa-tester` — Run targeted regression tests on the affected system
+- Delegate to `producer` — Approve deployment timing and communication plan
 
 All three must return APPROVE before proceeding. If any returns CONCERNS or REJECT, do not deploy — surface the issue and resolve it first.
 

.agents/skills/resume-from-handoff/SKILL.md

@@ -302,3 +302,7 @@ closeouts should read or refresh the saved `## Session Worklist` in `active.md`.
 - Does not perform a full artifact gap audit; that is `$project-stage-detect`.
 - Does not replace reading the handoff; it operationalizes the canonical handoff
   into an oriented, prioritized session worklist.
+
+## Ported metadata
+
+This is a Codex-native support skill. It has no upstream Claude skill equivalent.

.agents/skills/skill-improve/SKILL.md

@@ -68,13 +68,17 @@ Read the full skill file at `.agents/skills/[name]/SKILL.md`.
 
 For each failing or warning **static** check, identify the exact gap:
 
-- **Check 1 fail** → which frontmatter field is missing
+- **Check 1 fail** → which Codex frontmatter field or required Ported-metadata
+  field/classification is missing, or which unsupported field is in frontmatter
 - **Check 2 fail** → how many phases found vs. minimum required
 - **Check 3 fail** → no verdict keywords anywhere in the skill body
-- **Check 4 fail** → Write or Edit in allowed-tools but no ask-before-write language
+- **Check 4 fail** → effective Write or Edit behavior from Ported metadata or a
+  Codex-native workflow has no ask-before-write language or declared exception
 - **Check 5 warn** → no follow-up or next-step section at the end
-- **Check 6 warn** → `context: fork` set but fewer than 5 phases found
-- **Check 7 warn** → argument-hint is empty or doesn't match documented modes
+- **Check 6 warn** → declared Codex delegation lacks central authorization,
+  review-mode routing, missing-delegation reporting, or no-simulation fallback
+- **Check 7 warn** → Ported-metadata argument-hint is empty or doesn't match
+  documented modes; Codex-native argument-free support skills are not warned
 
 For each failing or warning **category** check (if category was assigned in Phase 2b),
 identify the exact gap in the skill's text. For example:

.agents/skills/skill-test/SKILL.md

@@ -39,15 +39,24 @@ If argument is missing or unrecognized, output usage and stop.
 
 For each skill being tested, read its `SKILL.md` fully and run all 7 checks:
 
-### Check 1 — Required Frontmatter Fields
-The file must contain all of these in the YAML frontmatter block:
+### Check 1 — Codex Metadata Contract
+The YAML frontmatter block must contain only the Codex runtime fields required
+for discovery:
 - `name:`
 - `description:`
+
+For a ported upstream skill, the `## Ported metadata` section must preserve:
 - `argument-hint:`
 - `user-invocable:`
 - `allowed-tools:`
 
-**FAIL** if any are absent.
+Codex-native support skills with no upstream equivalent instead must state that
+classification explicitly in `## Ported metadata`; they do not invent legacy
+fields merely to satisfy this check.
+
+**FAIL** if `name` or `description` is absent, if unsupported Claude fields are
+placed in YAML frontmatter, or if a ported skill omits its required Ported
+metadata fields.
 
 ### Check 2 — Multiple Phases
 The skill must have ≥2 numbered phase headings. Look for patterns like:
@@ -74,7 +83,12 @@ when they are explicitly described as derived checkpoints and say: Do not ask a
 separate "May I write?" for this file.
 
 **WARN** if absent (some read-only skills legitimately skip this).
-**FAIL** if `allowed-tools` includes `Write` or `Edit` but no ask-before-write language is found.
+For ported skills, read `allowed-tools` from the Ported metadata section, not
+from YAML frontmatter. For Codex-native support skills, infer write behavior
+from the workflow's declared actions and explicit invocation exceptions.
+
+**FAIL** if the effective tool/actions contract includes Write or Edit but no
+ask-before-write or valid declared exception language is found.
 
 ### Check 5 — Next-Step Handoff
 The skill must end with a recommended next action or follow-up path. Look for:
@@ -113,17 +127,21 @@ Skills that create, update, append, overwrite, or write
 checkpoint and must say: Do not ask a separate "May I write?" for this file.
 Durable artifact writes still require normal explicit approval.
 
-### Check 6 — Fork Context Complexity
-If frontmatter contains `context: fork`, the skill should have ≥5 phase headings
-(`##` level or numbered Phase N headers). Fork context is for complex multi-phase
-skills; simple skills should not use it.
+### Check 6 — Codex Delegation Structure
+If the skill declares role-agent or director delegation, verify that it uses the
+central `AGENTS.md` delegation authorization, applies review-mode filtering when
+director gates are involved, reports unavailable/skipped/blocked delegation,
+and never simulates a specialist or director verdict.
 
-**WARN** if `context: fork` is set but fewer than 5 phases found.
+**WARN** if declared delegation lacks one of those Codex-native routing or
+fallback requirements. Non-delegating skills pass this check as not applicable.
 
 ### Check 7 — Argument Hint Plausibility
-`argument-hint` must be non-empty. If the skill body mentions multiple modes
-(e.g., "Mode A | Mode B"), the hint should reflect them. Cross-reference the
-hint against the first phase's "Parse Arguments" section.
+For ported skills, the Ported-metadata `argument-hint` must be non-empty. If the
+skill body mentions multiple modes (e.g., "Mode A | Mode B"), the hint should
+reflect them. Cross-reference the hint against the first phase's "Parse
+Arguments" section. Codex-native support skills without an upstream hint pass
+when their usage is unambiguous or explicitly argument-free.
 
 **WARN** if hint is `""` or if documented modes don't match hint.
 
@@ -135,12 +153,12 @@ For a single skill:
 ```
 === Skill Static Check: /[name] ===
 
-Check 1 — Frontmatter Fields:    PASS
+Check 1 — Codex Metadata Contract: PASS
 Check 2 — Multiple Phases:       PASS (7 phases found)
 Check 3 — Verdict Keywords:      PASS (PASS, FAIL, CONCERNS)
 Check 4 — Collaborative Protocol: PASS ("May I write" found)
 Check 5 — Next-Step Handoff:     WARN (no follow-up section found)
-Check 6 — Fork Context Complexity: PASS (8 phases, context: fork set)
+Check 6 — Codex Delegation Structure: PASS (central routing and fallback present)
 Check 7 — Argument Hint:         PASS
 
 Verdict: WARNINGS (1 warning, 0 failures)

.agents/skills/studio-status/SKILL.md

@@ -28,3 +28,7 @@ context, current-directory, and git footer items remain configured separately
 through `[tui].status_line` when project config is installed. A real footer
 `Stage:` item is blocked until Codex documents support for project custom footer
 items; do not add an unverified fake status-line entry.
+
+## Ported metadata
+
+This is a Codex-native support skill. It has no upstream Claude skill equivalent.

.agents/skills/team-audio/SKILL.md

@@ -38,11 +38,11 @@ Store the resolved mode for use in all subsequent phases.
 ## How to Delegate
 
 Use the Codex subagent delegation to spawn each team member as a subagent:
-- `subagent_type: audio-director` — Sonic identity, emotional tone, audio palette
-- `subagent_type: sound-designer` — SFX specifications, audio events, mixing groups
-- `subagent_type: technical-artist` — Audio middleware, bus structure, memory budgets
-- `subagent_type: [primary engine specialist]` — Validate audio integration patterns for the engine
-- `subagent_type: gameplay-programmer` — Audio manager, gameplay triggers, adaptive music
+- Delegate to `audio-director` — Sonic identity, emotional tone, audio palette
+- Delegate to `sound-designer` — SFX specifications, audio events, mixing groups
+- Delegate to `technical-artist` — Audio middleware, bus structure, memory budgets
+- Delegate to `[primary engine specialist]` — Validate audio integration patterns for the engine
+- Delegate to `gameplay-programmer` — Audio manager, gameplay triggers, adaptive music
 
 Always provide full context in each agent's prompt (feature description, existing audio assets, design doc references).