Hide the three ambient skills from the slash menu; tighten the privacy response (#715)

A full HypAware install claimed eight slash-menu slots on Claude Code. Three of
those skills are ambient: they fire from natural phrasing and nobody types them
by name.

Mark `hypaware-query`, `hypaware-reference`, and `hypaware-graph`
`user-invocable: false`. They stay model-invocable, which is their only real
entry point. The explicit user commands keep their slots: `hypaware-ignore`,
`hypaware-unignore`, and `hypaware-privacy` are privacy controls that need a
guaranteed typed invocation rather than a phrasing that may or may not trip a
model trigger, and `hypaware-report` publishes to the org server.

Codex uses the same frontmatter key with the same spelling and semantics
(`user-invocable: false for background/reference-only skills`, alongside
`disable-model-invocation` and `allowed-tools`), so the key is set on both
client copies. That also holds the `deepEqual` frontmatter invariant in
test/plugins/query-skill-description.test.js and adds no new host divergence
for test/plugins/skill-host-parity.test.js, so the fixture is unchanged.
LLP 0075 is untouched: it records that Codex shares no *capture* signal with
Claude, not that Codex lacks a skills system.

Separately, hypaware-privacy grew a response-shape instruction at the end of
`## After the review`, a section describing what is true once the review is
done. Move it into Step 5, which already governs how findings are presented,
and keep the session-level purge path: `hyp purge --session <id>` covers a
sensitive session inside an otherwise-fine directory, which a directories-only
instruction would suppress.

Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>

Author: bgmcmullen

hypaware-core/plugins-workspace/claude/skills/hypaware-privacy/SKILL.md

@@ -129,6 +129,10 @@ Present findings as **short, redacted excerpts** and a proposed class per direct
 
 Then **apply nothing without per-item user confirmation.** Propose, wait for a yes on each item, then mark. Do not batch-apply.
 
+Keep the response tight: a clear list of candidate directories with the proposed
+class for each, minimal prose, no restating of the steps. Flag individual
+sessions separately only when a directory is otherwise fine but one session is not.
+
 ## Step 6 - Apply only via `hyp` verbs, and offer purge for every ignore (R6, R7)
 
 Apply each confirmed decision **only** through the `hyp` verbs below. **Never** author policy files or write anything into the user's repositories - the machine-local store is the only target.

hypaware-core/plugins-workspace/claude/skills/hypaware-query/SKILL.md

@@ -1,6 +1,7 @@
 ---
 name: hypaware-query
 description: Query this machine's recorded AI session history with the hyp query CLI. Covers every client HypAware records here, including Claude Code, Claude Desktop, Codex, OpenClaw, Hermes, and direct Anthropic/OpenAI API traffic. Use whenever the user refers to something they did before and the answer is not in the current conversation, even when they never say "HypAware", for example "what was I doing yesterday", "my most recent session", "which session did I work on X in", "have I hit this error before", "did I already try that", "what did that cost in tokens". Also use it to search recorded conversations for a topic, file, or repo, and for recorded logs, traces, metrics, AI gateway exchanges, query cache freshness, or SQL over local HypAware data. If you are about to grep or read ~/.claude/projects or ~/.codex/sessions, use this instead. For connections between sessions, files, and tools use hypaware-graph; for team-wide usage reporting use hypaware-report.
+user-invocable: false
 ---
 
 # HypAware Query

hypaware-core/plugins-workspace/claude/skills/hypaware-reference/SKILL.md

@@ -1,6 +1,7 @@
 ---
 name: hypaware-reference
 description: Explain what HypAware is, what it captures, how its data flows, config and paths, joining a fleet, and what is local-only versus opt-in. Use for product orientation - "what is HypAware", "what can it capture", "how do I detach codex", "how do I join a server", "where does my data go". For querying recorded data use hypaware-query; for graph questions hypaware-graph; for team token analysis hypaware-report.
+user-invocable: false
 ---
 
 # HypAware Reference

hypaware-core/plugins-workspace/codex/skills/hypaware-privacy/SKILL.md

@@ -221,6 +221,10 @@ Present findings as **short, redacted excerpts** and a proposed class per direct
 
 Then **apply nothing without per-item user confirmation.** Propose, wait for a yes on each item, then mark. Do not batch-apply.
 
+Keep the response tight: a clear list of candidate directories with the proposed
+class for each, minimal prose, no restating of the steps. Flag individual
+sessions separately only when a directory is otherwise fine but one session is not.
+
 ## Step 6 - Apply only via `hyp` verbs, and offer purge for every ignore (R6, R7)
 
 Apply each confirmed decision **only** through the `hyp` verbs below. **Never** author policy files or write anything into the user's repositories - the machine-local store is the only target.

hypaware-core/plugins-workspace/codex/skills/hypaware-query/SKILL.md

@@ -1,6 +1,7 @@
 ---
 name: hypaware-query
 description: Query this machine's recorded AI session history with the hyp query CLI. Covers every client HypAware records here, including Claude Code, Claude Desktop, Codex, OpenClaw, Hermes, and direct Anthropic/OpenAI API traffic. Use whenever the user refers to something they did before and the answer is not in the current conversation, even when they never say "HypAware", for example "what was I doing yesterday", "my most recent session", "which session did I work on X in", "have I hit this error before", "did I already try that", "what did that cost in tokens". Also use it to search recorded conversations for a topic, file, or repo, and for recorded logs, traces, metrics, AI gateway exchanges, query cache freshness, or SQL over local HypAware data. If you are about to grep or read ~/.claude/projects or ~/.codex/sessions, use this instead. For connections between sessions, files, and tools use hypaware-graph; for team-wide usage reporting use hypaware-report.
+user-invocable: false
 ---
 
 # HypAware Query

hypaware-core/plugins-workspace/codex/skills/hypaware-reference/SKILL.md

@@ -1,6 +1,7 @@
 ---
 name: hypaware-reference
 description: Explain what HypAware is, what it captures, how its data flows, config and paths, joining a fleet, and what is local-only versus opt-in. Use for product orientation - "what is HypAware", "what can it capture", "how do I detach codex", "how do I join a server", "where does my data go". For querying recorded data use hypaware-query; for graph questions hypaware-graph; for team token analysis hypaware-report.
+user-invocable: false
 ---
 
 # HypAware Reference

hypaware-core/plugins-workspace/context-graph/skills/hypaware-graph/SKILL.md

@@ -1,6 +1,7 @@
 ---
 name: hypaware-graph
 description: Explore the HypAware context graph: the activity graph projected from ai_gateway_messages (Sessions, Apps, Models, Tools, Files, Skills, Programs, and git Repos/Commits, and how they connect), plus optional server-side GitHub enrichment (Repos, PullRequests, Commits, Reviewers) that bridges AI sessions to code review. Use when the user asks what connects to a file/session/tool/model, which sessions ran a skill or invoked a program, wants co-occurrence or N-hop traversal, wants to join AI sessions to GitHub repos/PRs/reviewers, or wants to build/refresh the graph. Covers `hyp graph project` and `hyp graph neighbors`.
+user-invocable: false
 ---
 
 # HypAware Graph