crewly 1.1.2 → 1.2.0

package/config/skills/agent/competitor-content-tracker/instructions.md

@@ -0,0 +1,210 @@
+# Competitor Content Tracker
+
+Track and compare competitor content activity using browser automation and web search. Store structured data for trend analysis and content strategy.
+
+## Architecture
+
+1. **Browser scanning** (agent-driven) — Visit competitor pages, extract content data
+2. **Data management** (execute.sh) — Save, query, and compare stored data
+
+## Data Actions (execute.sh)
+
+### `save` — Store content items from a competitor scan
+
+| Parameter | Required | Description |
+|-----------|----------|-------------|
+| `action` | Yes | `"save"` |
+| `competitor` | Yes | `crewai`, `n8n`, `relevance-ai`, `autogen`, `langchain`, `langraph`, `openai`, `other` |
+| `sourceType` | Yes | `blog`, `twitter`, `linkedin`, `github-release`, `changelog`, `youtube`, `community`, `press`, `other` |
+| `items` | Yes | JSON array of content items (see schema below) |
+
+**Content item schema:**
+```json
+{
+  "title": "Blog post or content title",
+  "url": "https://...",
+  "publishedDate": "2026-02-27",
+  "description": "Brief summary",
+  "engagement": "45 likes, 12 comments",
+  "contentType": "blog post / tweet / video",
+  "tags": ["mcp", "agents", "enterprise"]
+}
+```
+
+### `list` — List tracked content files
+
+| Parameter | Required | Description |
+|-----------|----------|-------------|
+| `action` | Yes | `"list"` |
+| `competitor` | No | Filter by competitor |
+| `limit` | No | Max results (default: 20) |
+
+### `latest` — Get latest content from a competitor
+
+| Parameter | Required | Description |
+|-----------|----------|-------------|
+| `action` | Yes | `"latest"` |
+| `competitor` | Yes | Which competitor |
+| `sourceType` | No | Filter by source type |
+| `limit` | No | Max items (default: 15) |
+
+### `compare` — Compare content activity across competitors
+
+| Parameter | Required | Description |
+|-----------|----------|-------------|
+| `action` | Yes | `"compare"` |
+| `competitors` | No | Comma-separated list (default: `crewai,n8n,relevance-ai`) |
+| `days` | No | Lookback period in days (default: 7) |
+
+## Browser Scanning Guide
+
+### Competitor: CrewAI
+
+**Blog:**
+- URL: `https://www.crewai.com/blog`
+- Extract: title, date, summary, URL for each post
+- Focus on: product announcements, enterprise features, case studies
+
+**GitHub:**
+- URL: `https://github.com/crewAIInc/crewAI/releases`
+- Extract: version, release date, key changes
+- Also check: `https://docs.crewai.com/en/changelog`
+
+**Twitter/X:**
+- URL: `https://x.com/craborai` (CrewAI official)
+- Extract: recent tweets, engagement (likes, retweets, replies)
+- Focus on: product announcements, partnership news
+
+**PyPI (version tracking):**
+- URL: `https://pypi.org/project/crewai/#history`
+- Extract: version numbers, release dates
+
+**Example save:**
+```bash
+bash execute.sh '{"action":"save","competitor":"crewai","sourceType":"blog","projectPath":"/path","items":[
+  {"title":"The State of Agentic AI in 2026","url":"https://www.crewai.com/blog/state-of-agentic-ai","publishedDate":"2026-02-25","description":"Survey of 500 enterprises on AI agent adoption","engagement":"high - featured on front page","tags":["enterprise","survey","market-data"]},
+  {"title":"CrewAI v1.10.0 Release","url":"https://github.com/crewAIInc/crewAI/releases/tag/v1.10.0","publishedDate":"2026-02-27","description":"MCP enhancements, HITL improvements","engagement":"64 open issues","tags":["release","mcp","hitl"]}
+]}'
+```
+
+---
+
+### Competitor: n8n
+
+**Blog / Community:**
+- URL: `https://community.n8n.io/c/announcements/`
+- URL: `https://blog.n8n.io/`
+- Extract: announcements, feature updates, security bulletins
+
+**GitHub:**
+- URL: `https://github.com/n8n-io/n8n/releases`
+- Extract: version, date, key changes
+- Also: `https://docs.n8n.io/release-notes/`
+
+**Twitter/X:**
+- URL: `https://x.com/n8n_io`
+- Extract: recent tweets, engagement
+
+**Security (critical to track):**
+- URL: `https://community.n8n.io/tag/security`
+- Extract: CVE disclosures, patch versions
+
+**Example save:**
+```bash
+bash execute.sh '{"action":"save","competitor":"n8n","sourceType":"community","projectPath":"/path","items":[
+  {"title":"Security Bulletin February 25, 2026","url":"https://community.n8n.io/t/security-bulletin-february-25-2026/270324","publishedDate":"2026-02-25","description":"8 critical/high CVEs disclosed","engagement":"community discussion active","tags":["security","cve","critical"]}
+]}'
+```
+
+---
+
+### Competitor: Relevance AI
+
+**Changelog:**
+- URL: `https://relevanceai.com/changelog`
+- Extract: feature name, date, description
+
+**Blog:**
+- URL: `https://relevanceai.com/blog`
+- Extract: posts, topics, publication dates
+
+**Twitter/X:**
+- URL: `https://x.com/RelevanceAI_`
+- Extract: tweets, engagement
+
+---
+
+### Competitor: AutoGen (Microsoft)
+
+**GitHub:**
+- URL: `https://github.com/microsoft/autogen/releases`
+- Extract: version, changes, community engagement
+
+**Blog:**
+- URL: `https://microsoft.github.io/autogen/blog/`
+- Extract: posts, research papers
+
+---
+
+### Competitor: LangChain / LangGraph
+
+**Blog:**
+- URL: `https://blog.langchain.dev/`
+- Extract: posts, product announcements
+
+**GitHub:**
+- URL: `https://github.com/langchain-ai/langgraph/releases`
+- Extract: releases, features
+
+**Twitter/X:**
+- URL: `https://x.com/LangChainAI`
+
+---
+
+## Recommended Scan Schedule
+
+| Frequency | What to Scan |
+|-----------|-------------|
+| **Daily** | CrewAI Twitter, n8n Twitter — quick check for announcements |
+| **Twice/week** | CrewAI blog, n8n community, Relevance AI changelog |
+| **Weekly** | GitHub releases (all competitors), full comparison report |
+| **Ad-hoc** | When WebSearch reveals breaking news (acquisitions, security events, major launches) |
+
+## Workflow Integration
+
+```
+1. Agent scans competitor pages via browser/WebSearch
+2. Agent saves structured data via execute.sh save
+3. Agent runs compare to see cross-competitor activity
+4. Agent identifies content opportunities (reactive or counter-narrative)
+5. Agent feeds opportunities to trend-monitor suggest or content-writer draft
+```
+
+## Content Strategy Angles by Competitor
+
+### vs CrewAI
+- **Their strength:** Enterprise narrative, large community, Python ecosystem
+- **Our angle:** TypeScript/runtime-agnostic, live terminal, simpler setup for SMBs
+- **Content trigger:** When they release enterprise features, we publish "here's how SMBs get the same power without the complexity"
+
+### vs n8n
+- **Their strength:** Massive community (177K stars), mature integrations, visual workflow builder
+- **Our angle:** Security (PTY isolation vs their CVE history), AI-native (they added AI), full autonomy vs workflow steps
+- **Content trigger:** Security events, workflow builder limitations for AI tasks
+
+### vs Relevance AI
+- **Their strength:** No-code agents, marketplace scale (215+ agents)
+- **Our angle:** Developer control, open source, custom skills, not locked into their platform
+- **Content trigger:** When they push no-code, we publish "why developers need code-level control for AI agents"
+
+## Compare Report Example
+
+```bash
+# Weekly comparison across top 3 competitors
+bash execute.sh '{"action":"compare","competitors":"crewai,n8n,relevance-ai","days":"7","projectPath":"/path"}'
+```
+
+This outputs total items tracked, breakdown by source type, and recent titles for each competitor. Use this to identify:
+- Who is publishing most actively?
+- What topics are they focusing on?
+- Where are there gaps we can fill?

package/config/skills/agent/competitor-content-tracker/skill.json

@@ -0,0 +1,22 @@
+{
+  "id": "competitor-content-tracker",
+  "name": "Competitor Content Tracker",
+  "description": "Track and compare competitor content activity across blogs, social media, GitHub releases, and community channels. Supports CrewAI, n8n, Relevance AI, AutoGen, LangChain, and others.",
+  "category": "content",
+  "skillType": "claude-skill",
+  "promptFile": "instructions.md",
+  "execution": {
+    "type": "script",
+    "script": {
+      "file": "execute.sh",
+      "interpreter": "bash",
+      "timeoutMs": 15000
+    }
+  },
+  "assignableRoles": ["content-strategist", "product-manager", "generalist"],
+  "triggers": ["competitor content", "track competitor", "competitor analysis", "what are competitors posting", "competitive content"],
+  "tags": ["competitors", "content", "tracking", "analysis", "crewai", "n8n", "relevance-ai", "marketing", "competitive-intelligence"],
+  "version": "1.0.0",
+  "author": "Luna (Content Strategist)",
+  "license": "MIT"
+}

package/config/skills/agent/content-calendar/execute.sh

@@ -0,0 +1,294 @@
+#!/bin/bash
+# Content Calendar Manager — CRUD operations for content scheduling
+set -euo pipefail
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+source "${SCRIPT_DIR}/../_common/lib.sh"
+
+INPUT="${1:-}"
+[ -z "$INPUT" ] && error_exit "Usage: execute.sh '{\"action\":\"add|list|update|next|stats\",\"calendarPath\":\"/path/to/calendar.json\",...}'"
+
+# Parse common parameters
+ACTION=$(echo "$INPUT" | jq -r '.action // empty')
+CALENDAR_PATH=$(echo "$INPUT" | jq -r '.calendarPath // empty')
+
+require_param "action" "$ACTION"
+
+# Default calendar path
+if [ -z "$CALENDAR_PATH" ]; then
+  # Try to find project .crewly directory
+  PROJECT_PATH=$(echo "$INPUT" | jq -r '.projectPath // empty')
+  if [ -n "$PROJECT_PATH" ]; then
+    CALENDAR_PATH="${PROJECT_PATH}/.crewly/content/calendar.json"
+  else
+    CALENDAR_PATH="${HOME}/.crewly/content/calendar.json"
+  fi
+fi
+
+# Ensure calendar directory and file exist
+CALENDAR_DIR=$(dirname "$CALENDAR_PATH")
+mkdir -p "$CALENDAR_DIR"
+if [ ! -f "$CALENDAR_PATH" ]; then
+  echo '{"entries":[],"metadata":{"createdAt":"'"$(date -u +%Y-%m-%dT%H:%M:%SZ)"'","version":"1.0"}}' > "$CALENDAR_PATH"
+fi
+
+# Validate calendar JSON
+if ! jq empty "$CALENDAR_PATH" 2>/dev/null; then
+  error_exit "Calendar file is not valid JSON: $CALENDAR_PATH"
+fi
+
+case "$ACTION" in
+
+  # ─────────────────────────────────────────────
+  # ADD: Add a new content entry
+  # ─────────────────────────────────────────────
+  add)
+    TITLE=$(echo "$INPUT" | jq -r '.title // empty')
+    PLATFORM=$(echo "$INPUT" | jq -r '.platform // empty')
+    CONTENT_TYPE=$(echo "$INPUT" | jq -r '.type // "post"')
+    SCHEDULED_DATE=$(echo "$INPUT" | jq -r '.scheduledDate // empty')
+    STATUS=$(echo "$INPUT" | jq -r '.status // "draft"')
+    CONTENT_PATH=$(echo "$INPUT" | jq -r '.contentPath // empty')
+    LINE=$(echo "$INPUT" | jq -r '.line // "crewly"')
+    TOPIC=$(echo "$INPUT" | jq -r '.topic // empty')
+    NOTES=$(echo "$INPUT" | jq -r '.notes // empty')
+    TAGS=$(echo "$INPUT" | jq -r '.tags // "[]"')
+
+    require_param "title" "$TITLE"
+    require_param "platform" "$PLATFORM"
+    require_param "scheduledDate" "$SCHEDULED_DATE"
+
+    # Validate platform
+    VALID_PLATFORMS="x|linkedin|xiaohongshu|substack|youtube|github|reddit"
+    if ! echo "$PLATFORM" | grep -qE "^(${VALID_PLATFORMS})$"; then
+      error_exit "Invalid platform: $PLATFORM. Valid: x, linkedin, xiaohongshu, substack, youtube, github, reddit"
+    fi
+
+    # Validate status
+    VALID_STATUSES="idea|draft|ready|in-review|approved|published|archived"
+    if ! echo "$STATUS" | grep -qE "^(${VALID_STATUSES})$"; then
+      error_exit "Invalid status: $STATUS. Valid: idea, draft, ready, in-review, approved, published, archived"
+    fi
+
+    # Validate content type
+    VALID_TYPES="post|thread|article|video|image-text|newsletter|showcase|tutorial"
+    if ! echo "$CONTENT_TYPE" | grep -qE "^(${VALID_TYPES})$"; then
+      error_exit "Invalid type: $CONTENT_TYPE. Valid: post, thread, article, video, image-text, newsletter, showcase, tutorial"
+    fi
+
+    # Generate entry ID (timestamp-based)
+    ENTRY_ID="cc-$(date +%s)-$((RANDOM % 1000))"
+    NOW=$(date -u +%Y-%m-%dT%H:%M:%SZ)
+
+    # Ensure tags is valid JSON array
+    if ! echo "$TAGS" | jq empty 2>/dev/null; then
+      TAGS="[]"
+    fi
+
+    # Build new entry
+    NEW_ENTRY=$(jq -n \
+      --arg id "$ENTRY_ID" \
+      --arg title "$TITLE" \
+      --arg platform "$PLATFORM" \
+      --arg type "$CONTENT_TYPE" \
+      --arg scheduledDate "$SCHEDULED_DATE" \
+      --arg status "$STATUS" \
+      --arg contentPath "$CONTENT_PATH" \
+      --arg line "$LINE" \
+      --arg topic "$TOPIC" \
+      --arg notes "$NOTES" \
+      --argjson tags "$TAGS" \
+      --arg createdAt "$NOW" \
+      --arg updatedAt "$NOW" \
+      '{
+        id: $id,
+        title: $title,
+        platform: $platform,
+        type: $type,
+        scheduledDate: $scheduledDate,
+        status: $status,
+        contentPath: $contentPath,
+        line: $line,
+        topic: $topic,
+        notes: $notes,
+        tags: $tags,
+        createdAt: $createdAt,
+        updatedAt: $updatedAt
+      }')
+
+    # Append to calendar
+    jq --argjson entry "$NEW_ENTRY" '.entries += [$entry]' "$CALENDAR_PATH" > "${CALENDAR_PATH}.tmp" \
+      && mv "${CALENDAR_PATH}.tmp" "$CALENDAR_PATH"
+
+    jq -n --argjson entry "$NEW_ENTRY" '{"success":true,"action":"add","entry":$entry}'
+    ;;
+
+  # ─────────────────────────────────────────────
+  # LIST: List calendar entries with filters
+  # ─────────────────────────────────────────────
+  list)
+    FILTER_PLATFORM=$(echo "$INPUT" | jq -r '.platform // empty')
+    FILTER_STATUS=$(echo "$INPUT" | jq -r '.status // empty')
+    FILTER_DATE=$(echo "$INPUT" | jq -r '.date // empty')
+    FILTER_DATE_FROM=$(echo "$INPUT" | jq -r '.dateFrom // empty')
+    FILTER_DATE_TO=$(echo "$INPUT" | jq -r '.dateTo // empty')
+    FILTER_LINE=$(echo "$INPUT" | jq -r '.line // empty')
+    LIMIT=$(echo "$INPUT" | jq -r '.limit // "50"')
+
+    # Build jq filter
+    JQ_FILTER=".entries"
+
+    if [ -n "$FILTER_PLATFORM" ]; then
+      JQ_FILTER="${JQ_FILTER} | map(select(.platform == \"${FILTER_PLATFORM}\"))"
+    fi
+    if [ -n "$FILTER_STATUS" ]; then
+      JQ_FILTER="${JQ_FILTER} | map(select(.status == \"${FILTER_STATUS}\"))"
+    fi
+    if [ -n "$FILTER_DATE" ]; then
+      JQ_FILTER="${JQ_FILTER} | map(select(.scheduledDate == \"${FILTER_DATE}\"))"
+    fi
+    if [ -n "$FILTER_DATE_FROM" ]; then
+      JQ_FILTER="${JQ_FILTER} | map(select(.scheduledDate >= \"${FILTER_DATE_FROM}\"))"
+    fi
+    if [ -n "$FILTER_DATE_TO" ]; then
+      JQ_FILTER="${JQ_FILTER} | map(select(.scheduledDate <= \"${FILTER_DATE_TO}\"))"
+    fi
+    if [ -n "$FILTER_LINE" ]; then
+      JQ_FILTER="${JQ_FILTER} | map(select(.line == \"${FILTER_LINE}\"))"
+    fi
+
+    # Sort by scheduled date and limit
+    JQ_FILTER="${JQ_FILTER} | sort_by(.scheduledDate) | .[:${LIMIT}]"
+
+    ENTRIES=$(jq "$JQ_FILTER" "$CALENDAR_PATH")
+    COUNT=$(echo "$ENTRIES" | jq 'length')
+
+    jq -n \
+      --argjson entries "$ENTRIES" \
+      --argjson count "$COUNT" \
+      --arg calendarPath "$CALENDAR_PATH" \
+      '{"success":true,"action":"list","count":$count,"entries":$entries,"calendarPath":$calendarPath}'
+    ;;
+
+  # ─────────────────────────────────────────────
+  # UPDATE: Update an existing entry
+  # ─────────────────────────────────────────────
+  update)
+    ENTRY_ID=$(echo "$INPUT" | jq -r '.id // empty')
+    require_param "id" "$ENTRY_ID"
+
+    # Check if entry exists
+    EXISTS=$(jq --arg id "$ENTRY_ID" '[.entries[] | select(.id == $id)] | length' "$CALENDAR_PATH")
+    if [ "$EXISTS" -eq 0 ]; then
+      error_exit "Entry not found: $ENTRY_ID"
+    fi
+
+    NOW=$(date -u +%Y-%m-%dT%H:%M:%SZ)
+
+    # Build update fields dynamically
+    UPDATE_FIELDS=".updatedAt = \"${NOW}\""
+
+    NEW_STATUS=$(echo "$INPUT" | jq -r '.status // empty')
+    NEW_TITLE=$(echo "$INPUT" | jq -r '.title // empty')
+    NEW_DATE=$(echo "$INPUT" | jq -r '.scheduledDate // empty')
+    NEW_CONTENT_PATH=$(echo "$INPUT" | jq -r '.contentPath // empty')
+    NEW_NOTES=$(echo "$INPUT" | jq -r '.notes // empty')
+    NEW_TOPIC=$(echo "$INPUT" | jq -r '.topic // empty')
+
+    [ -n "$NEW_STATUS" ] && UPDATE_FIELDS="${UPDATE_FIELDS} | .status = \"${NEW_STATUS}\""
+    [ -n "$NEW_TITLE" ] && UPDATE_FIELDS="${UPDATE_FIELDS} | .title = \"${NEW_TITLE}\""
+    [ -n "$NEW_DATE" ] && UPDATE_FIELDS="${UPDATE_FIELDS} | .scheduledDate = \"${NEW_DATE}\""
+    [ -n "$NEW_CONTENT_PATH" ] && UPDATE_FIELDS="${UPDATE_FIELDS} | .contentPath = \"${NEW_CONTENT_PATH}\""
+    [ -n "$NEW_NOTES" ] && UPDATE_FIELDS="${UPDATE_FIELDS} | .notes = \"${NEW_NOTES}\""
+    [ -n "$NEW_TOPIC" ] && UPDATE_FIELDS="${UPDATE_FIELDS} | .topic = \"${NEW_TOPIC}\""
+
+    # If marking as published, record publishedAt
+    if [ "$NEW_STATUS" = "published" ]; then
+      UPDATE_FIELDS="${UPDATE_FIELDS} | .publishedAt = \"${NOW}\""
+    fi
+
+    # Apply update
+    jq --arg id "$ENTRY_ID" \
+      "(.entries[] | select(.id == \$id)) |= (${UPDATE_FIELDS})" \
+      "$CALENDAR_PATH" > "${CALENDAR_PATH}.tmp" \
+      && mv "${CALENDAR_PATH}.tmp" "$CALENDAR_PATH"
+
+    # Return updated entry
+    UPDATED=$(jq --arg id "$ENTRY_ID" '.entries[] | select(.id == $id)' "$CALENDAR_PATH")
+    jq -n --argjson entry "$UPDATED" '{"success":true,"action":"update","entry":$entry}'
+    ;;
+
+  # ─────────────────────────────────────────────
+  # NEXT: Get the next content to publish
+  # ─────────────────────────────────────────────
+  next)
+    FILTER_PLATFORM=$(echo "$INPUT" | jq -r '.platform // empty')
+    TODAY=$(date -u +%Y-%m-%d)
+
+    # Find entries that are ready/approved and scheduled for today or earlier
+    JQ_FILTER='.entries | map(select(.status == "ready" or .status == "approved"))'
+    JQ_FILTER="${JQ_FILTER} | map(select(.scheduledDate <= \"${TODAY}\"))"
+
+    if [ -n "$FILTER_PLATFORM" ]; then
+      JQ_FILTER="${JQ_FILTER} | map(select(.platform == \"${FILTER_PLATFORM}\"))"
+    fi
+
+    JQ_FILTER="${JQ_FILTER} | sort_by(.scheduledDate) | .[0] // null"
+
+    NEXT_ENTRY=$(jq "$JQ_FILTER" "$CALENDAR_PATH")
+
+    if [ "$NEXT_ENTRY" = "null" ]; then
+      # Also show upcoming entries
+      UPCOMING=$(jq --arg today "$TODAY" \
+        '[.entries[] | select(.status == "draft" or .status == "ready" or .status == "approved") | select(.scheduledDate > $today)] | sort_by(.scheduledDate) | .[:3]' \
+        "$CALENDAR_PATH")
+      jq -n --argjson upcoming "$UPCOMING" \
+        '{"success":true,"action":"next","next":null,"message":"No content ready to publish right now.","upcoming":$upcoming}'
+    else
+      jq -n --argjson entry "$NEXT_ENTRY" \
+        '{"success":true,"action":"next","next":$entry}'
+    fi
+    ;;
+
+  # ─────────────────────────────────────────────
+  # STATS: Calendar statistics
+  # ─────────────────────────────────────────────
+  stats)
+    TOTAL=$(jq '.entries | length' "$CALENDAR_PATH")
+    BY_STATUS=$(jq '[.entries[] | .status] | group_by(.) | map({(.[0]): length}) | add // {}' "$CALENDAR_PATH")
+    BY_PLATFORM=$(jq '[.entries[] | .platform] | group_by(.) | map({(.[0]): length}) | add // {}' "$CALENDAR_PATH")
+    BY_LINE=$(jq '[.entries[] | .line] | group_by(.) | map({(.[0]): length}) | add // {}' "$CALENDAR_PATH")
+
+    TODAY=$(date -u +%Y-%m-%d)
+    OVERDUE=$(jq --arg today "$TODAY" \
+      '[.entries[] | select(.scheduledDate < $today and (.status == "draft" or .status == "ready" or .status == "approved"))] | length' \
+      "$CALENDAR_PATH")
+    THIS_WEEK_END=$(date -u -v+7d +%Y-%m-%d 2>/dev/null || date -u -d "+7 days" +%Y-%m-%d 2>/dev/null || echo "$TODAY")
+    THIS_WEEK=$(jq --arg from "$TODAY" --arg to "$THIS_WEEK_END" \
+      '[.entries[] | select(.scheduledDate >= $from and .scheduledDate <= $to)] | length' \
+      "$CALENDAR_PATH")
+
+    jq -n \
+      --argjson total "$TOTAL" \
+      --argjson byStatus "$BY_STATUS" \
+      --argjson byPlatform "$BY_PLATFORM" \
+      --argjson byLine "$BY_LINE" \
+      --argjson overdue "$OVERDUE" \
+      --argjson thisWeek "$THIS_WEEK" \
+      --arg calendarPath "$CALENDAR_PATH" \
+      '{
+        "success": true,
+        "action": "stats",
+        "total": $total,
+        "byStatus": $byStatus,
+        "byPlatform": $byPlatform,
+        "byLine": $byLine,
+        "overdue": $overdue,
+        "thisWeek": $thisWeek,
+        "calendarPath": $calendarPath
+      }'
+    ;;
+
+  *)
+    error_exit "Unknown action: $ACTION. Valid actions: add, list, update, next, stats"
+    ;;
+esac

package/config/skills/agent/content-calendar/instructions.md

@@ -0,0 +1,122 @@
+# Content Calendar Manager
+
+Manage a content calendar with CRUD operations. Track planned content across platforms, schedule dates, and publication status.
+
+## Storage
+
+Calendar data is stored as a JSON file. Default location: `{projectPath}/.crewly/content/calendar.json`
+
+## Actions
+
+### `add` — Add a new content entry
+
+| Parameter | Required | Description |
+|-----------|----------|-------------|
+| `action` | Yes | `"add"` |
+| `title` | Yes | Content title or headline |
+| `platform` | Yes | `x`, `linkedin`, `xiaohongshu`, `substack`, `youtube`, `github`, `reddit` |
+| `type` | No | `post`, `thread`, `article`, `video`, `image-text`, `newsletter`, `showcase`, `tutorial` (default: `post`) |
+| `scheduledDate` | Yes | Target publish date in `YYYY-MM-DD` format |
+| `status` | No | `idea`, `draft`, `ready`, `in-review`, `approved`, `published`, `archived` (default: `draft`) |
+| `contentPath` | No | Path to the content file (markdown) |
+| `line` | No | Content line: `crewly` (brand) or `personal` (Steve). Default: `crewly` |
+| `topic` | No | Topic or brief description |
+| `notes` | No | Internal notes |
+| `tags` | No | JSON array of tags, e.g. `["ai-agent","security"]` |
+| `projectPath` | No | Project path to determine calendar location |
+| `calendarPath` | No | Custom calendar file path (overrides default) |
+
+### `list` — List entries with filters
+
+| Parameter | Required | Description |
+|-----------|----------|-------------|
+| `action` | Yes | `"list"` |
+| `platform` | No | Filter by platform |
+| `status` | No | Filter by status |
+| `date` | No | Filter by exact date (`YYYY-MM-DD`) |
+| `dateFrom` | No | Filter entries from this date |
+| `dateTo` | No | Filter entries up to this date |
+| `line` | No | Filter by content line (`crewly` or `personal`) |
+| `limit` | No | Max entries to return (default: 50) |
+
+### `update` — Update an entry
+
+| Parameter | Required | Description |
+|-----------|----------|-------------|
+| `action` | Yes | `"update"` |
+| `id` | Yes | Entry ID to update |
+| `status` | No | New status |
+| `title` | No | New title |
+| `scheduledDate` | No | New date |
+| `contentPath` | No | New content file path |
+| `notes` | No | New notes |
+| `topic` | No | New topic |
+
+### `next` — Get next content to publish
+
+Returns the next entry with status `ready` or `approved` that is scheduled for today or earlier.
+
+| Parameter | Required | Description |
+|-----------|----------|-------------|
+| `action` | Yes | `"next"` |
+| `platform` | No | Filter by platform |
+
+### `stats` — Calendar statistics
+
+Returns counts by status, platform, and content line. Also shows overdue and this-week counts.
+
+| Parameter | Required | Description |
+|-----------|----------|-------------|
+| `action` | Yes | `"stats"` |
+
+## Examples
+
+### Add a new X thread for Monday
+```bash
+bash execute.sh '{"action":"add","title":"AI Agent security - lessons from n8n CVEs","platform":"x","type":"thread","scheduledDate":"2026-03-03","status":"draft","line":"crewly","topic":"AI agent security","tags":["ai-agent","security","n8n"],"projectPath":"/path/to/project"}'
+```
+
+### List all content for next week
+```bash
+bash execute.sh '{"action":"list","dateFrom":"2026-03-03","dateTo":"2026-03-07","projectPath":"/path/to/project"}'
+```
+
+### List only draft X content
+```bash
+bash execute.sh '{"action":"list","platform":"x","status":"draft","projectPath":"/path/to/project"}'
+```
+
+### Mark content as published
+```bash
+bash execute.sh '{"action":"update","id":"cc-1709012345-42","status":"published","projectPath":"/path/to/project"}'
+```
+
+### Get next content to publish
+```bash
+bash execute.sh '{"action":"next","projectPath":"/path/to/project"}'
+```
+
+### Get calendar stats
+```bash
+bash execute.sh '{"action":"stats","projectPath":"/path/to/project"}'
+```
+
+## Status Workflow
+
+```
+idea → draft → ready → in-review → approved → published
+                                                  ↓
+                                              archived
+```
+
+- **idea**: Just a topic/concept, no content written
+- **draft**: Content being written
+- **ready**: Content complete, pending review
+- **in-review**: Sent to Steve for review
+- **approved**: Steve approved, ready to publish
+- **published**: Live on the platform
+- **archived**: Removed from active calendar
+
+## Output
+
+All actions return JSON with `success: true` and the relevant data. Errors return JSON to stderr with `error` field.

package/config/skills/agent/content-calendar/skill.json

@@ -0,0 +1,22 @@
+{
+  "id": "content-calendar",
+  "name": "Content Calendar Manager",
+  "description": "CRUD operations for content scheduling. Add, list, update, and query content calendar entries. Supports filtering by platform, status, date range, and content line (brand vs personal).",
+  "category": "content",
+  "skillType": "claude-skill",
+  "promptFile": "instructions.md",
+  "execution": {
+    "type": "script",
+    "script": {
+      "file": "execute.sh",
+      "interpreter": "bash",
+      "timeoutMs": 10000
+    }
+  },
+  "assignableRoles": ["content-strategist", "product-manager", "generalist"],
+  "triggers": ["content calendar", "schedule content", "content plan", "what to publish", "next post", "content schedule"],
+  "tags": ["content", "calendar", "scheduling", "planning", "marketing", "editorial"],
+  "version": "1.0.0",
+  "author": "Luna (Content Strategist)",
+  "license": "MIT"
+}