feat: lessons-learned loop — win signals + skill_edit auto-apply

Adds two new hook signal types:
- correction_free_streak: 5 consecutive UserPromptSubmits without a correction phrase
- clean_recovery: 3 clean PostToolUse events after a struggle signal
  (tool_error_loop / dead_end / retry_loop)

Both carry active_skills/active_agents payloads computed from a 10-event
activity ring, so ADAM can attribute wins to whichever skill was active
during the streak/recovery.

Promotes skill_edit to auto-apply under a strict gate (all required):
- conf >= 4 + cross-session evidence (existing rules)
- # Why cites a win-signal entry whose active_skills includes target
- diff append-only, +lines <= 30
- resulting SKILL.md size <= 2x current size
- 7-day cooldown per target (last_auto_edit in applied/ frontmatter)
- 30-day blacklist on user rejection (auto_apply_blacklist in rejected/)

Skill enforces the gate at apply time as defense in depth: re-stats target,
re-checks cooldown and blacklist, verifies append-only, reverts and refuses
on byte-cap breach. User-rejected skill_edit proposals automatically write
auto_apply_blacklist: true.

Win signals participate in the existing v0.2.0 source_entries archive
lifecycle, so already-applied evidence does not re-cluster.

Test suite: +5 cases (5 new asserts pass), 27 total passing.

Spec:  ~/.claude/docs/superpowers/specs/2026-05-10-adam-proactive-design.md
Plan:  ~/.claude/docs/superpowers/plans/2026-05-10-adam-proactive.md

adam/tests/run-tests.sh

@@ -279,6 +279,78 @@ else
 fi
 rm -rf /tmp/adam-test-18
 
+# --- Test 19: correction_free_streak fires after 5 clean prompts ---
+echo "Test 19: correction_free_streak after 5 clean prompts"
+reset_state
+for i in 1 2 3 4 5; do
+  echo "{\"hook_event_name\":\"UserPromptSubmit\",\"prompt\":\"please do step $i\",\"session_id\":\"sCF\",\"cwd\":\"/tmp/x\"}" \
+    | node "$HOOK" >/dev/null 2>&1 || true
+done
+assert_grep "$ROOT/journal.jsonl" '"type":"correction_free_streak"' "5 clean prompts logs correction_free_streak"
+
+# --- Test 20: correction phrase resets streak counter ---
+echo "Test 20: correction phrase breaks correction_free_streak"
+reset_state
+for i in 1 2 3 4; do
+  echo "{\"hook_event_name\":\"UserPromptSubmit\",\"prompt\":\"please do step $i\",\"session_id\":\"sCB\",\"cwd\":\"/tmp/x\"}" \
+    | node "$HOOK" >/dev/null 2>&1 || true
+done
+echo '{"hook_event_name":"UserPromptSubmit","prompt":"no, undo that","session_id":"sCB","cwd":"/tmp/x"}' \
+  | node "$HOOK" >/dev/null 2>&1 || true
+echo '{"hook_event_name":"UserPromptSubmit","prompt":"go on","session_id":"sCB","cwd":"/tmp/x"}' \
+  | node "$HOOK" >/dev/null 2>&1 || true
+if grep -qE '"type":"correction_free_streak"' "$ROOT/journal.jsonl"; then
+  echo "  FAIL: correction_free_streak fired despite intervening correction"; FAIL=$((FAIL+1))
+else
+  echo "  PASS: correction phrase reset the streak counter"; PASS=$((PASS+1))
+fi
+
+# --- Test 21: clean_recovery fires after struggle + 3 clean tools ---
+echo "Test 21: clean_recovery after struggle + 3 clean tools"
+reset_state
+for i in 1 2 3; do
+  echo '{"hook_event_name":"PostToolUse","tool_name":"Bash","tool_input":{"command":"foo"},"tool_response":{"is_error":true,"content":"Error: command not found: foo"},"session_id":"sR","cwd":"/tmp/x"}' \
+    | node "$HOOK" >/dev/null 2>&1 || true
+done
+for i in 1 2 3; do
+  echo "{\"hook_event_name\":\"PostToolUse\",\"tool_name\":\"Read\",\"tool_input\":{\"file_path\":\"/tmp/ok-$i\"},\"tool_response\":{\"content\":\"ok\"},\"session_id\":\"sR\",\"cwd\":\"/tmp/x\"}" \
+    | node "$HOOK" >/dev/null 2>&1 || true
+done
+assert_grep "$ROOT/journal.jsonl" '"type":"clean_recovery"' "3 clean tools after struggle logs clean_recovery"
+assert_grep "$ROOT/journal.jsonl" '"recovered_from":"tool_error_loop"' "recovered_from set on clean_recovery"
+
+# --- Test 22: clean_recovery resets when error breaks the streak ---
+echo "Test 22: clean_recovery suppressed by intervening error"
+reset_state
+for i in 1 2 3; do
+  echo '{"hook_event_name":"PostToolUse","tool_name":"Bash","tool_input":{"command":"foo"},"tool_response":{"is_error":true,"content":"Error: command not found: foo"},"session_id":"sRE","cwd":"/tmp/x"}' \
+    | node "$HOOK" >/dev/null 2>&1 || true
+done
+for i in 1 2; do
+  echo "{\"hook_event_name\":\"PostToolUse\",\"tool_name\":\"Read\",\"tool_input\":{\"file_path\":\"/tmp/ok-$i\"},\"tool_response\":{\"content\":\"ok\"},\"session_id\":\"sRE\",\"cwd\":\"/tmp/x\"}" \
+    | node "$HOOK" >/dev/null 2>&1 || true
+done
+echo '{"hook_event_name":"PostToolUse","tool_name":"Bash","tool_input":{"command":"x"},"tool_response":{"is_error":true,"content":"Error: again"},"session_id":"sRE","cwd":"/tmp/x"}' \
+  | node "$HOOK" >/dev/null 2>&1 || true
+echo '{"hook_event_name":"PostToolUse","tool_name":"Read","tool_input":{"file_path":"/tmp/ok-3"},"tool_response":{"content":"ok"},"session_id":"sRE","cwd":"/tmp/x"}' \
+  | node "$HOOK" >/dev/null 2>&1 || true
+if grep -qE '"type":"clean_recovery"' "$ROOT/journal.jsonl"; then
+  echo "  FAIL: clean_recovery fired despite intervening error"; FAIL=$((FAIL+1))
+else
+  echo "  PASS: clean_recovery suppressed by intervening error"; PASS=$((PASS+1))
+fi
+
+# --- Test 23: active_skills payload populated on win signals ---
+echo "Test 23: correction_free_streak payload includes active skill"
+reset_state
+echo '{"hook_event_name":"PreToolUse","tool_name":"Skill","tool_input":{"skill":"caveman"},"session_id":"sAS","cwd":"/tmp/x"}' \
+  | node "$HOOK" >/dev/null 2>&1 || true
+for i in 1 2 3 4 5; do
+  echo "{\"hook_event_name\":\"UserPromptSubmit\",\"prompt\":\"step $i\",\"session_id\":\"sAS\",\"cwd\":\"/tmp/x\"}" \
+    | node "$HOOK" >/dev/null 2>&1 || true
+done
+assert_grep "$ROOT/journal.jsonl" '"active_skills":\["caveman"\]' "active_skills payload includes invoked skill"
+
 echo
 echo "Results: $PASS passed, $FAIL failed"
 [ "$FAIL" = "0" ]

agents/adam.md

@@ -43,6 +43,8 @@ The hook emits these `type` values into the journal:
 | `edit_churn` | same file edited 4× in window | file basename |
 | `build_loop` | 2 build/test/compile commands fail in session | session |
 | `subagent_dispatch_pattern` | same subagent dispatched ≥3× cumulatively | subagent_type |
+| `correction_free_streak` | 5 clean UserPromptSubmits in a row (no correction phrase) | `active_skills[0]` |
+| `clean_recovery` | 3 clean PostToolUse events after a `tool_error_loop`/`dead_end`/`retry_loop` | (`recovered_from`, `active_skills[0]`) |
 
 ## Process
 
@@ -65,6 +67,8 @@ The hook emits these `type` values into the journal:
    - `edit_churn`: cluster by file basename pattern (e.g. `*.test.ts`).
    - `build_loop`: cluster by `session`.
    - `subagent_dispatch_pattern`: cluster by `subagent_type`.
+   - `correction_free_streak`: cluster by `active_skills[0]`. Treat ≥3 streaks across ≥2 sessions naming the same skill as cross-session evidence.
+   - `clean_recovery`: cluster by (`recovered_from`, `active_skills[0]`). A win cluster qualifies for `skill_edit` only when the named skill exists in `skills_root`.
 5. **Multi-axis correlation**: for each session that produced ≥2 distinct struggle types (`tool_error_loop`, `dead_end`, `weak_agent`, `retry_loop`, `edit_churn`, `build_loop`), tag clusters from that session as `multi_axis: true`. This grants +1 confidence at scoring.
 6. For each cluster qualifying under the rubric — ≥3 occurrences across ≥2 sessions, OR (for struggle types) ≥1 entry within a single session, OR (for `correction`) ≥3 occurrences across ≥2 cwds:
    a. If cluster topic matches a rejected idea via the rejected-ideas fuzzy set (≥2 token overlap with rejection's `# Why`), skip with reason `"rejected-similar"`.
@@ -172,6 +176,23 @@ Constraints:
 - Slug (used in `target` path filename) must not collide with any existing memory file.
 - For `type=feedback` and `type=project`, body MUST contain `**Why:**` and `**How to apply:**` lines (CLAUDE.md memory schema).
 
+## Win-driven `skill_edit` eligibility
+
+A `skill_edit` proposal sets `auto_apply_eligible: true` ONLY when ALL hold:
+
+1. `confidence ≥ 4`.
+2. `cross_session_evidence == true`.
+3. `# Why` cites ≥1 win-signal entry (`clean_recovery` or `correction_free_streak`) whose `active_skills` includes the target skill slug. Record this entry's `ts` in frontmatter field `win_evidence`.
+4. Diff is append-only — verify no `-` lines on existing SKILL.md content.
+5. Diff `+` lines ≤ 30.
+6. Resulting SKILL.md size ≤ 2× current size. Record both byte counts in frontmatter fields `bytes_before`, `bytes_after`.
+7. No entry in `applied_dir/` for the same `target` with `last_auto_edit` newer than 7 days ago (cooldown).
+8. No entry in `rejected_dir/` for this `target` with `auto_apply_blacklist: true` newer than 30 days ago.
+
+If any of (3)–(8) fails: still emit the proposal, but `auto_apply_eligible: false` — main thread queues for review.
+
+Win clusters do NOT override struggle clusters: a single `clean_recovery` cannot turn a `correction` cluster into a `skill_edit`. Struggle paths and win paths are independent.
+
 ## Confidence rubric (deterministic — do NOT vibe)
 
 Sum:
@@ -189,16 +210,16 @@ Sum:
 `auto_apply_eligible: true` requires **all** of:
 - `confidence ≥ 4`
 - `blast_radius == "low"`
-- `type ∈ {memory, skill_new}`
+- `type ∈ {memory, skill_new, skill_edit}` — `skill_edit` additionally requires the win-driven gate (see "Win-driven `skill_edit` eligibility")
 - `cross_session_evidence == true` — the +2 signal-repetition bonus came from the cross-session bullet (≥3× across ≥2 sessions). **Single-session-only struggle proposals always queue, never auto-apply, regardless of total confidence.** Record as frontmatter field `cross_session_evidence: true|false` on every proposal.
 
 ## Proposal types
 
 | Type | Target | Default blast | Auto-apply? |
 |---|---|---|---|
-| `memory` | `~/.claude/projects/<encoded-home>/memory/*.md` | low | yes if conf≥4 AND cross_session |
+| `memory` | `~/.claude/projects/-Users-nvm/memory/*.md` | low | yes if conf≥4 AND cross_session |
 | `skill_new` | new dir under `~/.claude/skills/` | low | yes if conf≥4 AND cross_session |
-| `skill_edit` | existing skill file | medium | no |
+| `skill_edit` | existing skill file | medium | yes if win-evidence + LOC + cooldown gates all pass (see "Win-driven skill_edit eligibility") |
 | `agent_new` | new file under `~/.claude/agents/` | medium | no |
 | `agent_edit` | existing agent file | medium | no |
 | `claude_md_edit` | `~/.claude/CLAUDE.md` | high | no |
@@ -243,6 +264,10 @@ source_entries:
   - "<journal entry ts that fed this cluster>"
   - "<another ts>"
   - "..."
+# skill_edit only — required when auto_apply_eligible: true
+win_evidence: "<ts of triggering clean_recovery or correction_free_streak entry>"
+bytes_before: <int>
+bytes_after: <int>
 ---
 
 # Why

hooks/adam-observe.mjs

@@ -28,6 +28,10 @@ const DEAD_END_THRESHOLD = 8;
 const EDIT_CHURN_THRESHOLD = 4;
 const BUILD_LOOP_THRESHOLD = 2;
 const SUBAGENT_DISPATCH_THRESHOLD = 3;
+const CORRECTION_FREE_THRESHOLD = 5;
+const CLEAN_RECOVERY_WINDOW = 3;
+const STRUGGLE_TYPES = new Set(["tool_error_loop", "dead_end", "retry_loop"]);
+const ACTIVE_SKILLS_LOOKBACK = 10;
 const STATE_MAX_BYTES = 1_000_000;
 
 function safeRead(path, fallback) {
@@ -77,6 +81,17 @@ function readUsage(name) {
   return usage[name] || 0;
 }
 
+function pushActivity(state, kind, name, ts) {
+  state.activity_ring.push({ kind, name, ts });
+  if (state.activity_ring.length > ACTIVE_SKILLS_LOOKBACK) state.activity_ring.shift();
+}
+
+function activeNames(state, kind) {
+  const seen = new Set();
+  for (const e of state.activity_ring) if (e.kind === kind) seen.add(e.name);
+  return [...seen];
+}
+
 function errorFingerprint(toolResponse) {
   if (!toolResponse) return null;
   let text = "";
@@ -114,6 +129,8 @@ function resetSessionLocal(state) {
   resetFrictionCounters(state);
   state.session_subagents = {};
   state.subagent_dispatch_emitted = {};
+  state.correctionFreeCounter = 0;
+  state.recoveryWatch = null;
 }
 
 function ensureStateDefaults(state) {
@@ -127,6 +144,9 @@ function ensureStateDefaults(state) {
   if (typeof state.build_loop_emitted !== "boolean") state.build_loop_emitted = false;
   if (!state.session_subagents || typeof state.session_subagents !== "object") state.session_subagents = {};
   if (!state.subagent_dispatch_emitted || typeof state.subagent_dispatch_emitted !== "object") state.subagent_dispatch_emitted = {};
+  if (typeof state.correctionFreeCounter !== "number") state.correctionFreeCounter = 0;
+  if (state.recoveryWatch === undefined) state.recoveryWatch = null;
+  if (!Array.isArray(state.activity_ring)) state.activity_ring = [];
 }
 
 function main() {
@@ -155,6 +175,18 @@ function main() {
         prev_tool: last.tool || null,
         prev_file: last.file || null,
       });
+      state.correctionFreeCounter = 0;
+    } else {
+      state.correctionFreeCounter += 1;
+      if (state.correctionFreeCounter >= CORRECTION_FREE_THRESHOLD) {
+        appendJournal({
+          ts, session, cwd, type: "correction_free_streak",
+          streak: state.correctionFreeCounter,
+          active_skills: activeNames(state, "skill"),
+          active_agents: activeNames(state, "agent"),
+        });
+        state.correctionFreeCounter = 0;
+      }
     }
     resetFrictionCounters(state);
   } else if (event === "PreToolUse") {
@@ -162,9 +194,11 @@ function main() {
     if (tool === "Skill") {
       const name = (input.tool_input && (input.tool_input.skill || input.tool_input.skill_name)) || "unknown";
       bumpUsage(`skill:${name}`);
+      pushActivity(state, "skill", name, ts);
     } else if (tool === "Agent") {
       const name = (input.tool_input && (input.tool_input.subagent_type || input.tool_input.agent)) || "unknown";
       bumpUsage(`agent:${name}`);
+      pushActivity(state, "agent", name, ts);
       state.session_subagents[name] = (state.session_subagents[name] || 0) + 1;
       const cumulative = readUsage(`agent:${name}`);
       const sessionCount = state.session_subagents[name];
@@ -182,6 +216,12 @@ function main() {
     const argsHash = djb2(JSON.stringify(input.tool_input || {}));
     const file = (input.tool_input && (input.tool_input.file_path || input.tool_input.path)) || null;
 
+    let struggleEmittedThisTurn = null;
+    const emit = (entry) => {
+      if (STRUGGLE_TYPES.has(entry.type)) struggleEmittedThisTurn = entry.type;
+      appendJournal(entry);
+    };
+
     const windowEntry = { tool, argsHash, file };
     if (tool === "Agent") {
       const sub = (input.tool_input && (input.tool_input.subagent_type || input.tool_input.agent)) || "unknown";
@@ -192,14 +232,14 @@ function main() {
 
     const sameToolArgs = state.tool_window.filter(e => e.tool === tool && e.argsHash === argsHash).length;
     if (sameToolArgs >= RETRY_THRESHOLD) {
-      appendJournal({ ts, session, cwd, type: "retry_loop", tool, count: sameToolArgs });
+      emit({ ts, session, cwd, type: "retry_loop", tool, count: sameToolArgs });
     }
 
     if (tool === "Agent") {
       const subagent = (input.tool_input && (input.tool_input.subagent_type || input.tool_input.agent)) || "unknown";
       const recent = state.tool_window.slice(-5).filter(e => e.tool === "Agent" && e.subagent === subagent).length;
       if (recent >= AGENT_RESPAWN_THRESHOLD) {
-        appendJournal({ ts, session, cwd, type: "weak_agent", subagent_type: subagent, count: recent });
+        emit({ ts, session, cwd, type: "weak_agent", subagent_type: subagent, count: recent });
       }
     }
 
@@ -214,14 +254,14 @@ function main() {
       if (state.last_errors.length > ERROR_RING_SIZE) state.last_errors.shift();
       const sameError = state.last_errors.filter(e => e.fp === fp).length;
       if (sameError >= ERROR_LOOP_THRESHOLD) {
-        appendJournal({ ts, session, cwd, type: "tool_error_loop", tool, count: sameError, fp });
+        emit({ ts, session, cwd, type: "tool_error_loop", tool, count: sameError, fp });
       }
     }
 
     if (file && EDIT_TOOLS.has(tool)) {
       state.edit_counts[file] = (state.edit_counts[file] || 0) + 1;
       if (state.edit_counts[file] >= EDIT_CHURN_THRESHOLD && !state.edit_churn_emitted[file]) {
-        appendJournal({ ts, session, cwd, type: "edit_churn", file, count: state.edit_counts[file] });
+        emit({ ts, session, cwd, type: "edit_churn", file, count: state.edit_counts[file] });
         state.edit_churn_emitted[file] = true;
       }
       const keys = Object.keys(state.edit_counts);
@@ -239,7 +279,7 @@ function main() {
       if (isBuildCmd && hasError) {
         state.build_failure_count += 1;
         if (state.build_failure_count >= BUILD_LOOP_THRESHOLD && !state.build_loop_emitted) {
-          appendJournal({ ts, session, cwd, type: "build_loop", count: state.build_failure_count, command: cmd.slice(0, 80) });
+          emit({ ts, session, cwd, type: "build_loop", count: state.build_failure_count, command: cmd.slice(0, 80) });
           state.build_loop_emitted = true;
         }
       }
@@ -247,9 +287,32 @@ function main() {
 
     state.tools_since_user += 1;
     if (state.tools_since_user >= DEAD_END_THRESHOLD && !state.dead_end_emitted) {
-      appendJournal({ ts, session, cwd, type: "dead_end", count: state.tools_since_user });
+      emit({ ts, session, cwd, type: "dead_end", count: state.tools_since_user });
       state.dead_end_emitted = true;
     }
+
+    if (struggleEmittedThisTurn) {
+      state.recoveryWatch = { recovered_from: struggleEmittedThisTurn, since_ts: ts, clean_count: 0, window_tools: [] };
+    } else if (state.recoveryWatch) {
+      const turnHadError = fp !== null;
+      if (turnHadError) {
+        state.recoveryWatch = null;
+      } else {
+        state.recoveryWatch.clean_count += 1;
+        state.recoveryWatch.window_tools.push(tool);
+        if (state.recoveryWatch.window_tools.length > CLEAN_RECOVERY_WINDOW) state.recoveryWatch.window_tools.shift();
+        if (state.recoveryWatch.clean_count >= CLEAN_RECOVERY_WINDOW) {
+          appendJournal({
+            ts, session, cwd, type: "clean_recovery",
+            recovered_from: state.recoveryWatch.recovered_from,
+            recovery_window_tools: state.recoveryWatch.window_tools.slice(),
+            active_skills: activeNames(state, "skill"),
+            active_agents: activeNames(state, "agent"),
+          });
+          state.recoveryWatch = null;
+        }
+      }
+    }
   }
 
   safeWrite(STATE, state);

skills/adam-self-improvement/SKILL.md

@@ -44,8 +44,22 @@ For each id in `high_confidence`:
 - Verify in front of the user: print `id`, `target`, `confidence`, `blast_radius`, `cross_session_evidence`, `auto_apply_eligible`.
 - Apply the change:
   - **For `skill_new`**: `mkdir -p ~/.claude/skills/<slug>/`, then `Write` the proposal's `# Proposed change` body to `~/.claude/skills/<slug>/SKILL.md`. After write, print: "skill `<slug>` written to `~/.claude/skills/<slug>/SKILL.md` — activates immediately — Claude Code v2.1.0+ auto-hot-reloads user-level skills, no restart needed."
-  - **For `memory`**: `Write` the proposal's `# Proposed change` body (which MUST include the auto-memory frontmatter — see "Memory drafting protocol" in `agents/adam.md`) to the path in `target` (under `~/.claude/projects/<encoded-home>/memory/`, where `<encoded-home>` is the user's home dir with `/` replaced by `-`, e.g. `-Users-alice` on macOS). Then update `MEMORY.md` index with a one-line pointer.
-  - **For other types under auto-apply**: apply via Write/Edit per `# Proposed change`. (Note: only `memory` and `skill_new` qualify for auto-apply per the rubric.)
+  - **For `memory`**: `Write` the proposal's `# Proposed change` body (which MUST include the auto-memory frontmatter — see "Memory drafting protocol" in `agents/adam.md`) to the path in `target`. Then update `MEMORY.md` index with a one-line pointer.
+  - **For `skill_edit`**: enforce the apply-time gate before writing.
+    1. Verify proposal frontmatter has `auto_apply_eligible: true`. If not, abort and queue for review.
+    2. Read `target` SKILL.md, capture `current_bytes` from a fresh stat — do NOT trust frontmatter `bytes_before`.
+    3. Verify diff in `# Proposed change`:
+       - Unified-diff format.
+       - Zero `-` lines on existing SKILL.md content (additions only).
+       - Total `+` lines ≤ 30.
+       If any check fails, print one-line refusal reason, leave proposal in `proposals/`, continue.
+    4. Cooldown re-check: scan `applied/` frontmatter for `target` matching this and `last_auto_edit` newer than 7 days ago. Refuse if found.
+    5. Blacklist re-check: scan `rejected/` frontmatter for `target` matching this and `auto_apply_blacklist: true` newer than 30 days ago. Refuse if found.
+    6. Apply via `Edit` tool (append the new section per the diff). Never use `Write` on existing SKILL.md.
+    7. Re-stat target. If new size exceeds `2 * current_bytes` (captured in step 2), revert via `Edit` (remove the just-appended section) and refuse — print refusal reason.
+    8. Add `last_auto_edit: <iso8601 utc now>` to the proposal frontmatter before moving it.
+    9. Tell user: "skill `<slug>` extended (added <N> lines) — auto-applied via win-evidence gate."
+  - **For other types under auto-apply**: apply via Write/Edit per `# Proposed change`. (Note: only `memory`, `skill_new`, and `skill_edit` qualify for auto-apply per the rubric.)
 - Move proposal to `~/.claude/adam/applied/<UTC-ts>-<id>.md`.
 - **Archive consumed journal entries**: `node ~/.claude/adam/scripts/adam-archive.mjs ~/.claude/adam/applied/<UTC-ts>-<id>.md` — moves entries listed in proposal's `source_entries` from `journal.jsonl` to `journal/actioned-<id>.jsonl` so subsequent `/reflect` runs do not re-cluster them.
 
@@ -66,7 +80,7 @@ c. On **approve**:
    - For all others: apply via Write/Edit per the proposal's `# Proposed change`.
    - Move proposal to `~/.claude/adam/applied/<ts>-<id>.md`.
    - Archive: `node ~/.claude/adam/scripts/adam-archive.mjs ~/.claude/adam/applied/<ts>-<id>.md`.
-d. On **reject**: ask for reason in one line. Append `# Reason\n<reason>` to proposal body. Move to `~/.claude/adam/rejected/<id>.md`. Archive: `node ~/.claude/adam/scripts/adam-archive.mjs ~/.claude/adam/rejected/<id>.md`.
+d. On **reject**: ask for reason in one line. Append `# Reason\n<reason>` to proposal body. If the proposal `type` is `skill_edit`, ALSO add `auto_apply_blacklist: true` to its frontmatter (so future reflects skip auto-apply on this target for 30 days). Move to `~/.claude/adam/rejected/<id>.md`. Archive: `node ~/.claude/adam/scripts/adam-archive.mjs ~/.claude/adam/rejected/<id>.md`.
 e. On **edit**: ask the user for the change, edit the proposal in place, then loop back to step 3a for that same id.
 
 ### 4. Handle failures
@@ -96,7 +110,7 @@ Before writing any proposal:
 - For `claude_md_edit`: confirm 3+ distinct cwds in the `# Why` section.
 - For `deletion`: confirm both criteria (a) and (b) from the agent's special handling are documented in the proposal.
 - For `skill_new`: confirm the slug doesn't collide with any existing skill in `~/.claude/skills/`. If it does, refuse and ask user to rename.
-- For `skill_edit`: confirm the diff is append-only (no `-` lines that remove existing content) and that target SKILL.md exists.
+- For `skill_edit`: confirm the diff is append-only (no `-` lines that remove existing content) and that target SKILL.md exists. When auto-applying, ALSO re-verify the eligibility gate steps in §2 (cooldown, blacklist, byte cap) before any `Edit` call — never trust frontmatter alone.
 - For `memory`: confirm `# Proposed change` body starts with `---` frontmatter containing required fields `name`, `description`, `type`, `originSessionId`. Refuse if frontmatter missing — agent must redraft per the Memory drafting protocol.
 - Confirm `source_entries` is present in proposal frontmatter as a non-empty list (used for archive). Warn (do not refuse) if missing — legacy proposals from before v0.2.0 won't have it.