refactor: drop superpowers references in SKILL.md, add .claude/skills/ discovery

The skills are functionally complete and standalone — superpowers
references were attribution-only. Removed so a reader without the
superpowers plugin loaded isn't sent chasing dead links.

Plus: .claude/skills/<name> symlinks so Claude Code's Skill tool can
invoke them by name (analogous to the existing .claude/agents/
symlinks). Discovery section in skills/README.md updated to cover
both symlink sets.

.claude/skills/audit

@@ -0,0 +1 @@
+../../skills/audit

.claude/skills/brainstorm

@@ -0,0 +1 @@
+../../skills/brainstorm

.claude/skills/debug

@@ -0,0 +1 @@
+../../skills/debug

.claude/skills/implement

@@ -0,0 +1 @@
+../../skills/implement

.claude/skills/plan

@@ -0,0 +1 @@
+../../skills/plan

skills/README.md

@@ -59,16 +59,28 @@ plus a system-prompt body. The `description` field is the one-sentence
 summary the orchestrator (or a skill) reads to decide whether to
 dispatch.
 
-## Subagent-type discovery
+## Discovery
 
-Subagent dispatch needs the agent files reachable under
-`.claude/agents/`. The repo provides one symlink per skill (tracked
-in git, no setup needed after clone):
+Two symlink sets, both tracked in git, no setup needed after clone:
+
+**Skills** are reachable under `.claude/skills/` so Claude Code's
+`Skill` tool can invoke them:
 
 ```
-.claude/agents/implement -> skills/implement/agents
-.claude/agents/audit     -> skills/audit/agents
-.claude/agents/debug     -> skills/debug/agents
+.claude/skills/brainstorm -> skills/brainstorm
+.claude/skills/plan       -> skills/plan
+.claude/skills/implement  -> skills/implement
+.claude/skills/audit      -> skills/audit
+.claude/skills/debug      -> skills/debug
+```
+
+**Agents** are reachable under `.claude/agents/` so subagent
+dispatch can find them by `subagent_type`:
+
+```
+.claude/agents/implement  -> skills/implement/agents
+.claude/agents/audit      -> skills/audit/agents
+.claude/agents/debug      -> skills/debug/agents
 ```
 
 ## Conventions

skills/brainstorm/SKILL.md

@@ -16,8 +16,7 @@ the spec means shipping the user's intuition unfiltered, which is
 exactly the reactive-deference failure mode AILang's orchestrator
 discipline exists to prevent.
 
-This skill adapts `superpowers:brainstorming` to the AILang project,
-producing specs at `docs/specs/<milestone>.md`.
+It produces specs at `docs/specs/<milestone>.md`.
 
 ## When to Use / Skipping
 
@@ -207,8 +206,6 @@ Hand off carries:
 
 ## Cross-references
 
-- **Upstream pattern:** `superpowers:brainstorming` — source of
-  the hard-gate, sectioned design, user-review gate.
 - **Project source:** `docs/DESIGN.md` "Feature-acceptance
   criterion" — gate this skill applies during spec writing.
 - **Output target:** `skills/plan/SKILL.md` — only valid next

skills/debug/SKILL.md

@@ -36,9 +36,9 @@ NO FIXES WITHOUT ROOT CAUSE INVESTIGATION FIRST
 NO FIX WITHOUT A FAILING TEST FIRST
 ```
 
-Both clauses are non-negotiable. The first is from
-`superpowers:systematic-debugging`. The second is the project's
-own rule (formerly in CLAUDE.md "Bug fixes — TDD, always").
+Both clauses are non-negotiable. The second was formerly the
+CLAUDE.md "Bug fixes — TDD, always" section before the
+2026-05-09 skill-system migration.
 
 ## The Process — four phases
 
@@ -143,8 +143,6 @@ All of these mean: **return to Phase 1**.
 
 ## Cross-references
 
-- **Upstream pattern:** `superpowers:systematic-debugging` —
-  source of the four-phase Iron Law and Phase 4.5.
 - **Agent dispatched:** `skills/debug/agents/ailang-debugger.md` —
   carries the actual diagnostic work (read CLAUDE.md / DESIGN.md /
   JOURNAL, reproduce, write RED test, propose minimal fix).

skills/implement/SKILL.md

@@ -16,9 +16,6 @@ code quality) before moving on. Subagents have isolated context; the
 orchestrator only sees their reports. This is what gives AILang's
 development cycle its context budget back.
 
-The pattern is `superpowers:subagent-driven-development` adapted to
-the AILang agent toolchain.
-
 ## When to Use / Skipping
 
 Triggers:
@@ -181,10 +178,6 @@ Reviewer subagents typically need at least the standard model.
 
 ## Cross-references
 
-- **Upstream pattern:** `superpowers:subagent-driven-development` —
-  source of the per-task fresh-subagent + two-stage review pattern.
-  `superpowers:executing-plans` for the parallel-session variant
-  (rarely used in AILang).
 - **Agents dispatched:**
   - `skills/implement/agents/ailang-implementer.md` — task execution
   - `skills/implement/agents/ailang-tester.md` — E2E coverage

skills/plan/SKILL.md

@@ -15,9 +15,8 @@ that will be created or modified, and decomposes work into bites
 small enough that a subagent can execute one in 2-5 minutes
 without making judgement calls outside its remit.
 
-This skill adapts `superpowers:writing-plans` to the AILang
-project conventions: plans live at `docs/plans/<iteration>.md`,
-header references the parent spec, commit-per-task pattern
+Plans live at `docs/plans/<iteration>.md`, the header
+references the parent spec, and the commit-per-task pattern
 matches the existing JOURNAL style.
 
 ## When to Use / Skipping
@@ -45,7 +44,7 @@ EVERY STEP THAT CHANGES CODE INCLUDES THE CODE.
 EVERY STEP THAT RUNS A COMMAND INCLUDES THE COMMAND AND EXPECTED OUTPUT.
 ```
 
-Lifted verbatim from `superpowers:writing-plans`. Non-negotiable.
+Non-negotiable.
 
 ## The Process
 
@@ -206,8 +205,6 @@ task focus ("only Tasks 1-3 this run").
 
 ## Cross-references
 
-- **Upstream pattern:** `superpowers:writing-plans` — source of the
-  bite-sized + No-Placeholders rule.
 - **Input source:** `skills/brainstorm/SKILL.md` — produces the
   spec this skill consumes.
 - **Output target:** `skills/implement/SKILL.md` — runs the plan