crewly 1.14.1 → 1.16.0

package/config/skills/_common/lib.test.sh

@@ -1,7 +1,8 @@
 #!/bin/bash
 # =============================================================================
 # Tests for lib.sh — shared skills library
-# Covers: --file preprocessor, read_json_input, require_param, error_exit
+# Covers: --file preprocessor, read_json_input, require_param, error_exit,
+#         api_call output cap (CREWLY_SKILL_MAX_OUTPUT_BYTES)
 # =============================================================================
 set -euo pipefail
 
@@ -184,6 +185,72 @@ CREWLY_EOF
 RESULT=$(bash "$TEMP_DIR/test_skill.sh" --file "$TEMP_DIR/heredoc_mixed.json")
 assert_contains "heredoc handles mixed quotes" "it's fine" "$RESULT"
 
+# =============================================================================
+# api_call output cap (CREWLY_SKILL_MAX_OUTPUT_BYTES)
+# curl is mocked with a shell function that returns $MOCK_BODY + "\n200".
+# =============================================================================
+mkdir -p "$TEMP_DIR/skills/fake-skill"
+cat > "$TEMP_DIR/skills/fake-skill/execute.sh" << 'SKILL_EOF'
+#!/bin/bash
+set -euo pipefail
+source "$LIB_PATH"
+# Mock curl: emit the canned body followed by the http code line, like
+# `curl -w '\n%{http_code}'` does.
+curl() {
+  printf '%s\n%s' "$MOCK_BODY" "${MOCK_CODE:-200}"
+}
+api_call GET "/anything"
+SKILL_EOF
+chmod +x "$TEMP_DIR/skills/fake-skill/execute.sh"
+
+export CREWLY_HOME="$TEMP_DIR/crewly-home"
+CAP_DIR="$CREWLY_HOME/tmp/skill-output"
+
+# ---- Test 17: small body passes through untouched ----
+export MOCK_BODY='{"ok":true,"items":[1,2,3]}'
+RESULT=$(CREWLY_SKILL_MAX_OUTPUT_BYTES=100 bash "$TEMP_DIR/skills/fake-skill/execute.sh")
+assert_eq "api_call: body under cap passes through" '{"ok":true,"items":[1,2,3]}' "$RESULT"
+
+# ---- Test 18: oversized body becomes a valid JSON envelope ----
+BIG=$(jq -nc '{ok:true, rows:[range(0;400)|{id:., text:"row-\(.)-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"}]}')
+export MOCK_BODY="$BIG"
+RESULT=$(CREWLY_SKILL_MAX_OUTPUT_BYTES=1000 bash "$TEMP_DIR/skills/fake-skill/execute.sh")
+assert_eq "api_call: oversized body -> truncated=true" "true" "$(printf '%s' "$RESULT" | jq -r '.truncated')"
+assert_eq "api_call: envelope bytes == real byte count" "$(printf '%s' "$BIG" | LC_ALL=C wc -c | tr -d ' ')" "$(printf '%s' "$RESULT" | jq -r '.bytes')"
+assert_contains "api_call: envelope hint present" "pass --full or read the file" "$(printf '%s' "$RESULT" | jq -r '.hint')"
+assert_eq "api_call: head is the first 4000 chars" "${BIG:0:4000}" "$(printf '%s' "$RESULT" | jq -r '.head')"
+CAP_FILE=$(printf '%s' "$RESULT" | jq -r '.file')
+assert_contains "api_call: file lands under CREWLY_HOME/tmp/skill-output" "$CAP_DIR/fake-skill-" "$CAP_FILE"
+assert_eq "api_call: parked file holds the full body" "$BIG" "$(cat "$CAP_FILE")"
+
+# ---- Test 19: --full bypasses the cap (flag is detected at source time) ----
+RESULT=$(CREWLY_SKILL_MAX_OUTPUT_BYTES=1000 bash "$TEMP_DIR/skills/fake-skill/execute.sh" --full)
+assert_eq "api_call: --full returns the raw body" "$BIG" "$RESULT"
+
+# ---- Test 20: CREWLY_SKILL_FULL_OUTPUT=1 bypasses the cap ----
+RESULT=$(CREWLY_SKILL_MAX_OUTPUT_BYTES=1000 CREWLY_SKILL_FULL_OUTPUT=1 bash "$TEMP_DIR/skills/fake-skill/execute.sh")
+assert_eq "api_call: CREWLY_SKILL_FULL_OUTPUT=1 returns the raw body" "$BIG" "$RESULT"
+
+# ---- Test 21: cap of 0 disables truncation ----
+RESULT=$(CREWLY_SKILL_MAX_OUTPUT_BYTES=0 bash "$TEMP_DIR/skills/fake-skill/execute.sh")
+assert_eq "api_call: CREWLY_SKILL_MAX_OUTPUT_BYTES=0 disables the cap" "$BIG" "$RESULT"
+
+# ---- Test 22: parked files older than the TTL are pruned at the start of any api_call ----
+mkdir -p "$CAP_DIR"
+touch -t 202001010000 "$CAP_DIR/old-skill-stale.json"
+touch "$CAP_DIR/fresh-skill-recent.json"
+export MOCK_BODY='{"ok":true}'
+RESULT=$(CREWLY_SKILL_MAX_OUTPUT_BYTES=1000 bash "$TEMP_DIR/skills/fake-skill/execute.sh")
+assert_eq "api_call: prune runs even when the body is small" '{"ok":true}' "$RESULT"
+assert_eq "api_call: stale parked output is deleted" "missing" "$([ -f "$CAP_DIR/old-skill-stale.json" ] && echo present || echo missing)"
+assert_eq "api_call: fresh parked output is kept" "present" "$([ -f "$CAP_DIR/fresh-skill-recent.json" ] && echo present || echo missing)"
+
+# ---- Test 23: error responses are untouched by the cap ----
+export MOCK_BODY='{"error":"nope"}' MOCK_CODE=500
+RESULT=$(CREWLY_SKILL_MAX_OUTPUT_BYTES=5 bash "$TEMP_DIR/skills/fake-skill/execute.sh" 2>&1 || true)
+assert_contains "api_call: non-2xx still reports the error object" '"status":500' "$RESULT"
+unset MOCK_BODY MOCK_CODE CREWLY_HOME
+
 echo ""
 echo "=== Results: $PASS/$TOTAL passed, $FAIL failed ==="
 

package/config/skills/agent/core/count-json/SKILL.md

@@ -0,0 +1,61 @@
+---
+name: Count JSON
+description: Run a command that prints JSON and count the entries matching a field=value filter. Built for skill_output Key Results (e.g. "tickets in review") so a KR can be measured from any board or API that speaks JSON.
+version: 1.0.0
+category: measurement
+skillType: claude-skill
+assignableRoles:
+  - orchestrator
+  - team-leader
+  - developer
+  - devops
+  - operations
+  - ops
+triggers:
+  - count json
+  - measure kr
+tags:
+  - okr
+  - measurement
+  - json
+execution:
+  type: script
+  script:
+    file: execute.sh
+    interpreter: bash
+    timeoutMs: 60000
+---
+
+# Count JSON
+
+Runs a command, reads the JSON it prints (an array, or an object holding an
+array under `items` / `tickets` / `data`), applies an optional `field=value`
+filter, and prints `{"count": N, "total": M, "filter": "..."}`.
+
+Made for `measurementSource: "skill_output"` Key Results:
+
+```json
+{ "measurementConfig": { "skill": "count-json",
+                         "args": { "command": "cd /opt/app/web && make tickets TICKET_ARGS=--json",
+                                   "field": "status", "value": "review" },
+                         "jsonPath": ".count" } }
+```
+
+## Parameters
+
+| Name | Required | Meaning |
+|---|---|---|
+| `command` | yes | Shell command whose stdout is JSON (or a `file` path instead) |
+| `file` | no | Read JSON from this file instead of running a command |
+| `field` | no | Top-level field to filter on |
+| `value` | no | Value that `field` must equal (string compare) |
+| `arrayPath` | no | jq path to the array when it is nested (default: auto-detect) |
+
+## Output
+
+```json
+{"count": 6, "total": 21, "filter": "status=review"}
+```
+
+The command runs with a 50 s timeout; a non-JSON result exits 1 with an error
+object on stderr.

package/config/skills/agent/core/count-json/execute.sh

@@ -0,0 +1,52 @@
+#!/bin/bash
+# count-json — run a command (or read a file) that yields JSON, count entries
+# matching field=value. Output: {"count":N,"total":M,"filter":"field=value"}
+#
+# Usage: execute.sh '{"command":"...","field":"status","value":"review"}'
+#        execute.sh '{"file":"/tmp/board.json","field":"status","value":"open"}'
+set -uo pipefail
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+source "${SCRIPT_DIR}/../../_common/lib.sh"
+
+INPUT="${1:-}"
+[ -z "$INPUT" ] && error_exit "Usage: execute.sh '{\"command\":\"...\",\"field\":\"status\",\"value\":\"review\"}'"
+
+COMMAND=$(printf '%s' "$INPUT" | jq -r '.command // empty')
+FILE=$(printf '%s' "$INPUT" | jq -r '.file // empty')
+FIELD=$(printf '%s' "$INPUT" | jq -r '.field // empty')
+VALUE=$(printf '%s' "$INPUT" | jq -r '.value // empty')
+ARRAY_PATH=$(printf '%s' "$INPUT" | jq -r '.arrayPath // empty')
+TIMEOUT_S="${COUNT_JSON_TIMEOUT_S:-50}"
+
+[ -z "$COMMAND" ] && [ -z "$FILE" ] && error_exit "command or file is required"
+
+if [ -n "$FILE" ]; then
+  [ -f "$FILE" ] || error_exit "file not found: $FILE"
+  RAW=$(cat "$FILE")
+else
+  if command -v timeout >/dev/null 2>&1; then
+    RAW=$(timeout "$TIMEOUT_S" bash -c "$COMMAND" 2>/dev/null)
+  else
+    RAW=$(bash -c "$COMMAND" 2>/dev/null)
+  fi
+  [ $? -ne 0 ] && [ -z "$RAW" ] && error_exit "command failed or produced no output"
+fi
+
+# Skip any non-JSON preamble (progress lines, banners).
+JSON=$(printf '%s' "$RAW" | awk 'found||/^[[{]/{found=1; print}')
+[ -z "$JSON" ] && error_exit "no JSON found in output"
+
+if [ -n "$ARRAY_PATH" ]; then
+  SELECT="$ARRAY_PATH"
+else
+  SELECT='if type=="array" then . elif (.items|type)=="array" then .items elif (.tickets|type)=="array" then .tickets elif (.data|type)=="array" then .data else [] end'
+fi
+
+if [ -n "$FIELD" ]; then
+  printf '%s' "$JSON" | jq -c --arg f "$FIELD" --arg v "$VALUE" \
+    "($SELECT) as \$arr | {count: (\$arr | map(select((.[\$f] // \"\" | tostring) == \$v)) | length), total: (\$arr | length), filter: (\$f + \"=\" + \$v)}" \
+    || error_exit "output is not valid JSON"
+else
+  printf '%s' "$JSON" | jq -c "($SELECT) as \$arr | {count: (\$arr | length), total: (\$arr | length), filter: \"\"}" \
+    || error_exit "output is not valid JSON"
+fi

package/config/skills/agent/core/count-json/execute.test.sh

@@ -0,0 +1,35 @@
+#!/bin/bash
+# Tests for count-json — run with: bash execute.test.sh
+set -u
+HERE="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+EXEC="$HERE/execute.sh"
+PASS=0; FAIL=0
+check() {
+  local name="$1"; local got="$2"; local want="$3"
+  if [ "$got" = "$want" ]; then PASS=$((PASS+1)); else FAIL=$((FAIL+1)); echo "FAIL: $name"; echo "  want: $want"; echo "  got:  $got"; fi
+}
+run() { bash "$EXEC" "$1" 2>/dev/null; }
+run_err() { bash "$EXEC" "$1" 2>&1 >/dev/null | jq -r '.error' 2>/dev/null; }
+
+TMP=$(mktemp -d); trap 'rm -rf "$TMP"' EXIT
+printf '%s' '[{"ticketId":"T1","status":"open"},{"ticketId":"T2","status":"review"},{"ticketId":"T3","status":"review"},{"ticketId":"T4","status":"blocked"}]' > "$TMP/board.json"
+printf '%s' 'fetching...
+{"items":[{"s":"a"},{"s":"b"}]}' > "$TMP/nested.txt"
+
+ARGS=$(jq -cn --arg f "$TMP/board.json" '{file:$f, field:"status", value:"review"}')
+check "filter by field=value" "$(run "$ARGS")" '{"count":2,"total":4,"filter":"status=review"}'
+
+ARGS=$(jq -cn --arg f "$TMP/board.json" '{file:$f}')
+check "no filter counts all" "$(run "$ARGS")" '{"count":4,"total":4,"filter":""}'
+
+ARGS=$(jq -cn --arg c "cat $TMP/nested.txt" '{command:$c, field:"s", value:"b"}')
+check "command with preamble and nested items" "$(run "$ARGS")" '{"count":1,"total":2,"filter":"s=b"}'
+
+ARGS=$(jq -cn --arg f "$TMP/board.json" '{file:$f, arrayPath:".", field:"status", value:"open"}')
+check "explicit arrayPath" "$(run "$ARGS")" '{"count":1,"total":4,"filter":"status=open"}'
+
+check "non-JSON errors" "$(run_err '{"command":"echo not json"}')" "no JSON found in output"
+check "missing input errors" "$(run_err '{}')" "command or file is required"
+
+echo "=== Results: $PASS passed, $FAIL failed ==="
+[ "$FAIL" -eq 0 ]

package/config/skills/agent/core/record-prediction/SKILL.md

@@ -0,0 +1,61 @@
+---
+name: Record Prediction
+description: Record a prediction with a confidence level so your calibration can be measured once it resolves.
+version: 1.0.0
+category: memory
+skillType: claude-skill
+assignableRoles:
+  - developer
+  - qa
+  - tpm
+  - designer
+  - frontend-developer
+  - backend-developer
+  - fullstack-dev
+  - qa-engineer
+  - product-manager
+  - architect
+  - generalist
+  - sales
+  - support
+  - orchestrator
+triggers:
+  - record prediction
+  - I predict
+  - confidence
+  - estimate
+tags:
+  - memory
+  - calibration
+  - self-improvement
+execution:
+  type: script
+  script:
+    file: execute.sh
+    interpreter: bash
+    timeoutMs: 15000
+---
+
+# Record Prediction
+
+Write down a falsifiable prediction ("this PR lands today", "the flaky test is a race in the mock") with a confidence between 0 and 1. Later, close it with `resolve-prediction`. Your calibration score (how well confidence tracks reality) is shown in your prompt once predictions resolve, with guidance when you are over- or under-confident.
+
+## Parameters
+
+| Flag | JSON Field | Required | Description |
+|------|-----------|----------|-------------|
+| `--statement` / `-p` | `statement` | Yes | What you predict will happen |
+| `--confidence` / `-c` | `confidence` | Yes | 0–1 (e.g. `0.7`) |
+| `--resolve-by` | `resolveBy` | No | ISO date by which it should be resolved |
+| `--session` / `-s` | `sessionName` | No | Your session name — defaults to `$CREWLY_SESSION_NAME` |
+
+## Examples
+
+```bash
+bash execute.sh --statement "checkout PR merges before Friday" --confidence 0.7 --resolve-by 2026-10-03
+bash execute.sh '{"statement":"checkout PR merges before Friday","confidence":0.7}'
+```
+
+## Output
+
+`{ "success": true, "id": "pred-...", "prediction": {...} }` — keep the `id`; you need it to resolve.

package/config/skills/agent/core/record-prediction/execute.sh

@@ -0,0 +1,90 @@
+#!/bin/bash
+# =============================================================================
+# record-prediction — Record a prediction with a confidence level
+#
+# Backed by POST /api/agents/:sessionName/self-improvement/predictions.
+#
+# Usage:
+#   bash execute.sh --statement "PR lands today" --confidence 0.7 [--resolve-by 2026-10-01] [--session crewly-dev-1]
+#   bash execute.sh '{"statement":"PR lands today","confidence":0.7,"resolveBy":"2026-10-01"}'
+# =============================================================================
+set -euo pipefail
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+source "${SCRIPT_DIR}/../../_common/lib.sh"
+
+print_usage() {
+  cat <<'EOF_USAGE'
+Usage:
+  bash execute.sh --statement "PR lands today" --confidence 0.7 [--resolve-by 2026-10-01] [--session crewly-dev-1]
+  bash execute.sh '{"statement":"PR lands today","confidence":0.7}'
+
+Options:
+  --statement  | -p   What you predict (required)
+  --confidence | -c   Confidence 0-1 (required)
+  --resolve-by        ISO date the prediction should be resolved by (optional)
+  --session    | -s   Session name (defaults to $CREWLY_SESSION_NAME)
+  --help       | -h   Show this help
+EOF_USAGE
+}
+
+INPUT_JSON=""
+SESSION_NAME=""
+STATEMENT=""
+CONFIDENCE=""
+RESOLVE_BY=""
+
+if [[ $# -gt 0 && ${1:0:1} == '{' ]]; then
+  INPUT_JSON="$1"
+  shift || true
+fi
+
+while [[ $# -gt 0 ]]; do
+  case "$1" in
+    --statement|-p)
+      [ $# -ge 2 ] || error_exit "--statement requires a value"
+      STATEMENT="$2"; shift 2 ;;
+    --confidence|-c)
+      [ $# -ge 2 ] || error_exit "--confidence requires a value"
+      CONFIDENCE="$2"; shift 2 ;;
+    --resolve-by)
+      [ $# -ge 2 ] || error_exit "--resolve-by requires a value"
+      RESOLVE_BY="$2"; shift 2 ;;
+    --session|-s)
+      [ $# -ge 2 ] || error_exit "--session requires a value"
+      SESSION_NAME="$2"; shift 2 ;;
+    --help|-h) print_usage; exit 0 ;;
+    *) error_exit "Unknown option: $1" ;;
+  esac
+done
+
+if [ -n "$INPUT_JSON" ]; then
+  INPUT=$(read_json_input "$INPUT_JSON")
+  [ -z "$SESSION_NAME" ] && SESSION_NAME=$(printf '%s' "$INPUT" | jq -r '.sessionName // .agentId // empty')
+  [ -z "$STATEMENT" ] && STATEMENT=$(printf '%s' "$INPUT" | jq -r '.statement // .prediction // empty')
+  [ -z "$CONFIDENCE" ] && CONFIDENCE=$(printf '%s' "$INPUT" | jq -r 'if .confidence == null then empty else (.confidence | tostring) end')
+  [ -z "$RESOLVE_BY" ] && RESOLVE_BY=$(printf '%s' "$INPUT" | jq -r '.resolveBy // empty')
+fi
+
+[ -z "$SESSION_NAME" ] && SESSION_NAME="${CREWLY_SESSION_NAME:-}"
+require_param "sessionName (--session or CREWLY_SESSION_NAME)" "$SESSION_NAME"
+require_param "statement (--statement)" "$STATEMENT"
+require_param "confidence (--confidence)" "$CONFIDENCE"
+
+# Validate locally so a typo fails fast with a readable message.
+if ! printf '%s' "$CONFIDENCE" | grep -Eq '^(0(\.[0-9]+)?|1(\.0+)?)$'; then
+  error_exit "confidence must be a number between 0 and 1 (got: $CONFIDENCE)"
+fi
+
+if [ -n "$RESOLVE_BY" ]; then
+  BODY=$(jq -cn --arg s "$STATEMENT" --argjson c "$CONFIDENCE" --arg r "$RESOLVE_BY" '{statement: $s, confidence: $c, resolveBy: $r}')
+else
+  BODY=$(jq -cn --arg s "$STATEMENT" --argjson c "$CONFIDENCE" '{statement: $s, confidence: $c}')
+fi
+
+RESPONSE=$(api_call POST "/agents/${SESSION_NAME}/self-improvement/predictions" "$BODY" 2>&1) || {
+  ERROR_MSG=$(printf '%s' "$RESPONSE" | jq -r '.details.error // .details // .error // "Unknown error"' 2>/dev/null || echo "$RESPONSE")
+  jq -n --arg reason "$ERROR_MSG" '{success: false, reason: $reason}'
+  exit 1
+}
+
+printf '%s' "$RESPONSE" | jq -c '{success: (.success // false), id: (.data.prediction.id // null), prediction: (.data.prediction // null)}'

package/config/skills/agent/core/record-prediction/execute.test.sh

@@ -0,0 +1,67 @@
+#!/bin/bash
+# Tests for record-prediction — run with: bash execute.test.sh
+# Spins up a python HTTP stub as the backend, asserts the request path/body
+# the skill sends and the JSON it prints. Exit 0 on pass.
+set -u
+HERE="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+EXEC="$HERE/execute.sh"
+PASS=0; FAIL=0
+check() {
+  local name="$1"; local got="$2"; local want="$3"
+  if [ "$got" = "$want" ]; then PASS=$((PASS+1)); else FAIL=$((FAIL+1)); echo "FAIL: $name"; echo "  want: $want"; echo "  got:  $got"; fi
+}
+
+PORT=18793
+STUB_LOG="$(mktemp)"; export STUB_LOG; export STUB_PORT="$PORT"
+python3 -u <<'PY' >/dev/null 2>&1 &
+from http.server import BaseHTTPRequestHandler, HTTPServer
+import json, os
+LOG = os.environ['STUB_LOG']; PORT = int(os.environ['STUB_PORT'])
+class H(BaseHTTPRequestHandler):
+    def do_POST(self):
+        n = int(self.headers.get('content-length', '0')); body = self.rfile.read(n).decode() if n else ''
+        data = json.loads(body or '{}')
+        open(LOG, 'w').write(json.dumps({'path': self.path, 'body': data}))
+        if data.get('resolveBy') == 'soon':
+            self.send_response(400); self.send_header('Content-Type', 'application/json'); self.end_headers()
+            self.wfile.write(b'{"success":false,"error":"resolveBy must be an ISO date string"}'); return
+        pred = {'id': 'pred-1-abc', 'prediction': data.get('statement'), 'confidence': data.get('confidence'), 'madeAt': '2026-09-18'}
+        if 'resolveBy' in data: pred['resolveBy'] = data['resolveBy']
+        self.send_response(201); self.send_header('Content-Type', 'application/json'); self.end_headers()
+        self.wfile.write(json.dumps({'success': True, 'data': {'prediction': pred}}).encode())
+    def log_message(self, *a, **k): pass
+HTTPServer(('127.0.0.1', PORT), H).serve_forever()
+PY
+STUB_PID=$!
+disown "$STUB_PID" 2>/dev/null || true
+trap 'kill "$STUB_PID" >/dev/null 2>&1; rm -f "$STUB_LOG"' EXIT
+for i in $(seq 1 30); do curl -s -o /dev/null -m 0.2 "http://127.0.0.1:${PORT}/" 2>/dev/null && break; sleep 0.1; done
+
+run() { CREWLY_API_URL="http://127.0.0.1:${PORT}" CREWLY_SESSION_NAME="crewly-dev-1" bash "$EXEC" "$@" 2>/dev/null; }
+run_err() { CREWLY_API_URL="http://127.0.0.1:${PORT}" CREWLY_SESSION_NAME="" bash "$EXEC" "$@" 2>&1 >/dev/null | jq -r '.error' 2>/dev/null; }
+
+OUT=$(run --statement "PR lands today" --confidence 0.7 --resolve-by 2026-10-01)
+check "flags: output id" "$(printf '%s' "$OUT" | jq -r .id)" "pred-1-abc"
+check "flags: output success" "$(printf '%s' "$OUT" | jq -r .success)" "true"
+check "flags: path" "$(jq -r .path "$STUB_LOG")" "/api/agents/crewly-dev-1/self-improvement/predictions"
+check "flags: body (confidence is numeric, resolveBy present)" "$(jq -c .body "$STUB_LOG")" '{"statement":"PR lands today","confidence":0.7,"resolveBy":"2026-10-01"}'
+
+run '{"statement":"no deadline","confidence":1,"sessionName":"crewly-qa-9"}' >/dev/null
+check "json: body omits resolveBy when absent" "$(jq -c .body "$STUB_LOG")" '{"statement":"no deadline","confidence":1}'
+check "json: explicit session in path" "$(jq -r .path "$STUB_LOG")" "/api/agents/crewly-qa-9/self-improvement/predictions"
+
+run '{"prediction":"legacy field name","confidence":"0.25"}' >/dev/null
+check "json: accepts legacy 'prediction' key and string confidence" "$(jq -c .body "$STUB_LOG")" '{"statement":"legacy field name","confidence":0.25}'
+
+OUT=$(CREWLY_API_URL="http://127.0.0.1:${PORT}" CREWLY_SESSION_NAME="crewly-dev-1" bash "$EXEC" --statement x --confidence 0.5 --resolve-by soon 2>/dev/null); RC=$?
+check "backend 400: exit code" "$RC" "1"
+check "backend 400: reason" "$(printf '%s' "$OUT" | jq -r .reason)" "resolveBy must be an ISO date string"
+
+check "confidence > 1 rejected locally" "$(run_err --session s --statement x --confidence 1.5)" "confidence must be a number between 0 and 1 (got: 1.5)"
+check "non-numeric confidence rejected locally" "$(run_err --session s --statement x --confidence abc)" "confidence must be a number between 0 and 1 (got: abc)"
+check "missing statement errors" "$(run_err --session s --confidence 0.5)" "Missing required parameter: statement (--statement)"
+check "missing confidence errors" "$(run_err --session s --statement x)" "Missing required parameter: confidence (--confidence)"
+check "missing session errors" "$(run_err --statement x --confidence 0.5)" "Missing required parameter: sessionName (--session or CREWLY_SESSION_NAME)"
+
+echo "=== Results: $PASS passed, $FAIL failed ==="
+[ "$FAIL" -eq 0 ]

package/config/skills/agent/core/resolve-prediction/SKILL.md

@@ -0,0 +1,61 @@
+---
+name: Resolve Prediction
+description: Close a recorded prediction with what actually happened so your calibration score updates.
+version: 1.0.0
+category: memory
+skillType: claude-skill
+assignableRoles:
+  - developer
+  - qa
+  - tpm
+  - designer
+  - frontend-developer
+  - backend-developer
+  - fullstack-dev
+  - qa-engineer
+  - product-manager
+  - architect
+  - generalist
+  - sales
+  - support
+  - orchestrator
+triggers:
+  - resolve prediction
+  - prediction outcome
+  - was I right
+tags:
+  - memory
+  - calibration
+  - self-improvement
+execution:
+  type: script
+  script:
+    file: execute.sh
+    interpreter: bash
+    timeoutMs: 15000
+---
+
+# Resolve Prediction
+
+Record the outcome of a prediction made with `record-prediction`. Pass `--accurate true|false`; if you only pass an outcome, a plain verdict such as `correct`, `wrong`, `yes`, `no` is understood — anything else needs the explicit flag. The response includes your updated calibration score.
+
+## Parameters
+
+| Flag | JSON Field | Required | Description |
+|------|-----------|----------|-------------|
+| `--id` | `id` | Yes | Prediction id (`pred-...`) |
+| `--outcome` / `-o` | `outcome` | Yes | What actually happened |
+| `--accurate` / `-a` | `accurate` | No* | `true` / `false` — required unless outcome is a plain verdict |
+| `--session` / `-s` | `sessionName` | No | Your session name — defaults to `$CREWLY_SESSION_NAME` |
+
+## Examples
+
+```bash
+bash execute.sh --id pred-1718000000-ab12cd --outcome "merged Thursday" --accurate true
+bash execute.sh --id pred-1718000000-ab12cd --outcome wrong
+bash execute.sh '{"id":"pred-1718000000-ab12cd","outcome":"slipped a week","accurate":false}'
+```
+
+## Output
+
+`{ "success": true, "prediction": {...}, "calibrationScore": 0.82 }`

package/config/skills/agent/core/resolve-prediction/execute.sh

@@ -0,0 +1,87 @@
+#!/bin/bash
+# =============================================================================
+# resolve-prediction — Close a prediction with its actual outcome
+#
+# Backed by POST /api/agents/:sessionName/self-improvement/predictions/:id/resolve.
+#
+# Usage:
+#   bash execute.sh --id pred-123 --outcome "merged Thursday" --accurate true [--session crewly-dev-1]
+#   bash execute.sh --id pred-123 --outcome wrong
+#   bash execute.sh '{"id":"pred-123","outcome":"slipped","accurate":false}'
+# =============================================================================
+set -euo pipefail
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+source "${SCRIPT_DIR}/../../_common/lib.sh"
+
+print_usage() {
+  cat <<'EOF_USAGE'
+Usage:
+  bash execute.sh --id pred-123 --outcome "merged Thursday" --accurate true [--session crewly-dev-1]
+  bash execute.sh --id pred-123 --outcome wrong
+  bash execute.sh '{"id":"pred-123","outcome":"slipped","accurate":false}'
+
+Options:
+  --id               Prediction id (required)
+  --outcome  | -o    What actually happened (required)
+  --accurate | -a    true|false (required unless outcome is a plain verdict like "correct"/"wrong")
+  --session  | -s    Session name (defaults to $CREWLY_SESSION_NAME)
+  --help     | -h    Show this help
+EOF_USAGE
+}
+
+INPUT_JSON=""
+SESSION_NAME=""
+PRED_ID=""
+OUTCOME=""
+ACCURATE=""
+
+if [[ $# -gt 0 && ${1:0:1} == '{' ]]; then
+  INPUT_JSON="$1"
+  shift || true
+fi
+
+while [[ $# -gt 0 ]]; do
+  case "$1" in
+    --id)
+      [ $# -ge 2 ] || error_exit "--id requires a value"
+      PRED_ID="$2"; shift 2 ;;
+    --outcome|-o)
+      [ $# -ge 2 ] || error_exit "--outcome requires a value"
+      OUTCOME="$2"; shift 2 ;;
+    --accurate|-a)
+      [ $# -ge 2 ] || error_exit "--accurate requires a value"
+      ACCURATE="$2"; shift 2 ;;
+    --session|-s)
+      [ $# -ge 2 ] || error_exit "--session requires a value"
+      SESSION_NAME="$2"; shift 2 ;;
+    --help|-h) print_usage; exit 0 ;;
+    *) error_exit "Unknown option: $1" ;;
+  esac
+done
+
+if [ -n "$INPUT_JSON" ]; then
+  INPUT=$(read_json_input "$INPUT_JSON")
+  [ -z "$SESSION_NAME" ] && SESSION_NAME=$(printf '%s' "$INPUT" | jq -r '.sessionName // .agentId // empty')
+  [ -z "$PRED_ID" ] && PRED_ID=$(printf '%s' "$INPUT" | jq -r '.id // .predictionId // empty')
+  [ -z "$OUTCOME" ] && OUTCOME=$(printf '%s' "$INPUT" | jq -r '.outcome // empty')
+  [ -z "$ACCURATE" ] && ACCURATE=$(printf '%s' "$INPUT" | jq -r 'if .accurate == null then empty else (.accurate | tostring) end')
+fi
+
+[ -z "$SESSION_NAME" ] && SESSION_NAME="${CREWLY_SESSION_NAME:-}"
+require_param "sessionName (--session or CREWLY_SESSION_NAME)" "$SESSION_NAME"
+require_param "id (--id)" "$PRED_ID"
+require_param "outcome (--outcome)" "$OUTCOME"
+
+case "$ACCURATE" in
+  "") BODY=$(jq -cn --arg o "$OUTCOME" '{outcome: $o}') ;;
+  true|false) BODY=$(jq -cn --arg o "$OUTCOME" --argjson a "$ACCURATE" '{outcome: $o, accurate: $a}') ;;
+  *) error_exit "accurate must be true or false (got: $ACCURATE)" ;;
+esac
+
+RESPONSE=$(api_call POST "/agents/${SESSION_NAME}/self-improvement/predictions/${PRED_ID}/resolve" "$BODY" 2>&1) || {
+  ERROR_MSG=$(printf '%s' "$RESPONSE" | jq -r '.details.error // .details // .error // "Unknown error"' 2>/dev/null || echo "$RESPONSE")
+  jq -n --arg reason "$ERROR_MSG" '{success: false, reason: $reason}'
+  exit 1
+}
+
+printf '%s' "$RESPONSE" | jq -c '{success: (.success // false), prediction: (.data.prediction // null), calibrationScore: (.data.calibrationScore // null)}'