crewly 1.0.0

package/config/skills/agent/manage-knowledge/instructions.md

@@ -0,0 +1,46 @@
+# Manage Knowledge
+
+Create or update company knowledge documents. These are markdown files stored in the Crewly knowledge system, accessible to all team members via the UI.
+
+Use this skill to document SOPs, architecture decisions, team norms, onboarding guides, runbooks, and other company knowledge that should be shared across the team.
+
+**Note:** Agents can create and update documents but cannot delete them. Deletion is a user-only action via the UI.
+
+## Parameters
+
+| Parameter | Required | Description |
+|-----------|----------|-------------|
+| `action` | Yes | `create` or `update` |
+| `title` | Yes (create) | Document title |
+| `content` | Yes (create) | Markdown content |
+| `category` | No | Category: `SOPs`, `Team Norms`, `Architecture`, `Onboarding`, `Runbooks`, `General` (default: `General`) |
+| `scope` | No | `global` (company-wide) or `project` (project-specific). Default: `global` |
+| `projectPath` | Conditional | Required when scope is `project` |
+| `tags` | No | Array of tags for search/filtering |
+| `createdBy` | No | Your session name (default: `agent`) |
+| `id` | Yes (update) | Document ID to update |
+| `updatedBy` | No | Your session name for updates (default: `agent`) |
+
+## Examples
+
+### Create a global SOP
+
+```bash
+bash config/skills/agent/manage-knowledge/execute.sh '{"action":"create","title":"PR Review Process","content":"# PR Review Process\n\n1. All PRs require at least one approval\n2. Run tests locally before requesting review\n3. Keep PRs under 400 lines when possible","category":"SOPs","scope":"global","tags":["process","code-review"],"createdBy":"dev-1"}'
+```
+
+### Create a project-specific architecture doc
+
+```bash
+bash config/skills/agent/manage-knowledge/execute.sh '{"action":"create","title":"API Design Patterns","content":"# API Design Patterns\n\nWe use RESTful conventions with Express.js...","category":"Architecture","scope":"project","projectPath":"/path/to/project","tags":["api","backend"],"createdBy":"dev-1"}'
+```
+
+### Update an existing document
+
+```bash
+bash config/skills/agent/manage-knowledge/execute.sh '{"action":"update","id":"abc-123-uuid","content":"# Updated Content\n\nNew information added...","scope":"global","updatedBy":"dev-1"}'
+```
+
+## Output
+
+JSON response with `success: true` and the document ID on creation, or `success: true` on update.

package/config/skills/agent/nano-banana-image/.env

@@ -0,0 +1,2 @@
+# Nano Banana Image Generation API Configuration
+GEMINI_API_KEY=AIzaSyDWVSRxr0e9JkKBkq-Tkdy5fBEs0_HCJ-8

package/config/skills/agent/nano-banana-image/.env.example

@@ -0,0 +1,6 @@
+# Nano Banana Image Generation API Configuration
+# Copy this file to .env and add your API key
+
+# Google Gemini API Key (required)
+# Get your key at: https://aistudio.google.com/apikey
+GEMINI_API_KEY=your_api_key_here

package/config/skills/agent/nano-banana-image/generate.sh

@@ -0,0 +1,73 @@
+#!/bin/bash
+# Nano Banana Image Generation Script
+# This script calls the Gemini API to generate images
+# Supports both standard (flash) and pro models
+
+set -e
+
+# Check for required environment variable
+if [ -z "$GEMINI_API_KEY" ]; then
+    echo "Error: GEMINI_API_KEY environment variable is not set"
+    exit 1
+fi
+
+# Model selection
+# Usage: generate.sh [prompt] [--pro]
+#   --pro    Use gemini-3-pro-image-preview (higher quality)
+#   default  Use gemini-2.0-flash-exp-image-generation (faster)
+MODEL="gemini-2.0-flash-exp-image-generation"
+PROMPT=""
+
+for arg in "$@"; do
+    if [ "$arg" = "--pro" ]; then
+        MODEL="gemini-3-pro-image-preview"
+    elif [ -z "$PROMPT" ]; then
+        PROMPT="$arg"
+    fi
+done
+
+if [ -z "$PROMPT" ]; then
+    read -r PROMPT
+fi
+
+if [ -z "$PROMPT" ]; then
+    echo "Error: No prompt provided"
+    exit 1
+fi
+
+# API endpoint
+API_URL="https://generativelanguage.googleapis.com/v1beta/models/${MODEL}:generateContent"
+
+# Create the request payload
+PAYLOAD=$(cat <<EOF
+{
+  "contents": [
+    {
+      "parts": [
+        {
+          "text": "Generate an image: $PROMPT"
+        }
+      ]
+    }
+  ],
+  "generationConfig": {
+    "responseModalities": ["Text", "Image"]
+  }
+}
+EOF
+)
+
+# Make the API call
+RESPONSE=$(curl -s -X POST "$API_URL?key=$GEMINI_API_KEY" \
+    -H "Content-Type: application/json" \
+    -d "$PAYLOAD")
+
+# Check for errors
+if echo "$RESPONSE" | grep -q '"error"'; then
+    echo "API Error:"
+    echo "$RESPONSE" | jq -r '.error.message // .error'
+    exit 1
+fi
+
+# Output the response
+echo "$RESPONSE"

package/config/skills/agent/nano-banana-image/instructions.md

@@ -0,0 +1,50 @@
+# Nano Banana Image Generation
+
+This skill enables AI image generation through Google's Gemini API.
+
+## Overview
+
+The Nano Banana Image Generation skill allows agents to create images from text descriptions using Google's Gemini model. This is useful for:
+
+- Creating concept art and visualizations
+- Generating placeholder images for designs
+- Producing custom graphics based on descriptions
+- Rapid prototyping of visual content
+
+## Usage
+
+To generate an image, provide a detailed text description of what you want to create.
+
+### Example Prompts
+
+- "A futuristic cityscape at sunset with flying cars"
+- "A cozy coffee shop interior with warm lighting"
+- "An abstract representation of music and sound waves"
+
+## API Configuration
+
+This skill requires a Gemini API key to function. The API key should be stored in the skill's `.env` file:
+
+```
+GEMINI_API_KEY=your_api_key_here
+```
+
+### Getting an API Key
+
+1. Visit [Google AI Studio](https://aistudio.google.com/apikey)
+2. Sign in with your Google account
+3. Create a new API key
+4. Copy the key and add it to the skill configuration
+
+## Best Practices
+
+- Be specific and detailed in your image descriptions
+- Include style preferences (realistic, artistic, minimalist, etc.)
+- Specify desired colors, lighting, and mood
+- Mention any specific elements that must be included
+
+## Limitations
+
+- Image generation may take 30-60 seconds
+- Some content types may be restricted by the API
+- Results may vary based on prompt complexity

package/config/skills/agent/nano-banana-image/skill.json

@@ -0,0 +1,39 @@
+{
+  "id": "nano-banana-image",
+  "name": "Nano Banana Image Generation",
+  "description": "Generate images using Google's Gemini Nano Banana API. Supports standard (flash) and pro models. Use --pro flag for higher quality output with gemini-3-pro-image-preview.",
+  "category": "design",
+  "skillType": "claude-skill",
+  "promptFile": "instructions.md",
+  "execution": {
+    "type": "script",
+    "script": {
+      "file": "generate.sh",
+      "interpreter": "bash",
+      "timeoutMs": 120000
+    }
+  },
+  "environment": {
+    "file": ".env",
+    "required": ["GEMINI_API_KEY"]
+  },
+  "notices": [
+    {
+      "type": "requirement",
+      "title": "API Key Required",
+      "message": "This skill requires a Gemini API key. Add your API key in the skill settings.",
+      "link": "https://aistudio.google.com/apikey",
+      "linkText": "Get API Key"
+    }
+  ],
+  "triggers": ["generate image", "create image", "nano banana", "gemini image", "ai image"],
+  "models": {
+    "standard": "gemini-2.0-flash-exp-image-generation",
+    "pro": "gemini-3-pro-image-preview"
+  },
+  "tags": ["image", "generation", "ai", "gemini", "api", "pro"],
+  "assignableRoles": ["designer"],
+  "version": "1.1.0",
+  "createdAt": "2024-01-01T00:00:00.000Z",
+  "updatedAt": "2024-01-01T00:00:00.000Z"
+}

package/config/skills/agent/query-knowledge/execute.sh

@@ -0,0 +1,30 @@
+#!/bin/bash
+# Query Knowledge Documents - search company knowledge base
+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 '{\"query\":\"deployment\",\"scope\":\"global\"}'"
+
+QUERY=$(echo "$INPUT" | jq -r '.query // empty')
+SCOPE=$(echo "$INPUT" | jq -r '.scope // "global"')
+CATEGORY=$(echo "$INPUT" | jq -r '.category // empty')
+PROJECT_PATH=$(echo "$INPUT" | jq -r '.projectPath // empty')
+require_param "query" "$QUERY"
+
+# URL-encode the query parameters
+ENCODED_QUERY=$(printf '%s' "$QUERY" | jq -sRr @uri)
+ENDPOINT="/knowledge/documents?scope=${SCOPE}&search=${ENCODED_QUERY}"
+
+if [ -n "$CATEGORY" ]; then
+  ENCODED_CATEGORY=$(printf '%s' "$CATEGORY" | jq -sRr @uri)
+  ENDPOINT="${ENDPOINT}&category=${ENCODED_CATEGORY}"
+fi
+
+if [ -n "$PROJECT_PATH" ]; then
+  ENCODED_PATH=$(printf '%s' "$PROJECT_PATH" | jq -sRr @uri)
+  ENDPOINT="${ENDPOINT}&projectPath=${ENCODED_PATH}"
+fi
+
+api_call GET "$ENDPOINT"

package/config/skills/agent/query-knowledge/instructions.md

@@ -0,0 +1,47 @@
+# Query Knowledge
+
+Search the company knowledge base for relevant documents. Use this to find SOPs, runbooks, architecture docs, team norms, and other documentation before starting process-oriented tasks.
+
+Knowledge documents are stored at two scopes:
+- **global** (`~/.crewly/docs/`) — company-wide documents accessible to all agents
+- **project** (`{projectPath}/.crewly/docs/`) — project-specific documents
+
+## Parameters
+
+| Parameter | Required | Description |
+|-----------|----------|-------------|
+| `query` | Yes | Search text to match against document titles, tags, and content |
+| `scope` | No | `global` (default) or `project` |
+| `category` | No | Filter by category: `SOPs`, `Team Norms`, `Architecture`, `Onboarding`, `Runbooks`, `General` |
+| `projectPath` | Conditional | Required when scope is `project` |
+
+## Examples
+
+### Search global knowledge for deployment docs
+
+```bash
+bash {{AGENT_SKILLS_PATH}}/query-knowledge/execute.sh '{"query":"deployment","scope":"global"}'
+```
+
+### Search project-specific architecture docs
+
+```bash
+bash {{AGENT_SKILLS_PATH}}/query-knowledge/execute.sh '{"query":"API design","scope":"project","projectPath":"{{PROJECT_PATH}}","category":"Architecture"}'
+```
+
+### Search for runbooks
+
+```bash
+bash {{AGENT_SKILLS_PATH}}/query-knowledge/execute.sh '{"query":"incident response","scope":"global","category":"Runbooks"}'
+```
+
+## When to Use
+
+- **Before process-oriented tasks** — search for SOPs and runbooks that describe how to do things
+- **Before architecture work** — check for existing architecture docs and design decisions
+- **When onboarding to a project** — search for onboarding guides and team norms
+- **Before answering questions** about processes, conventions, or infrastructure
+
+## Output
+
+JSON response with `success: true` and a `data` array of matching document summaries (id, title, category, tags, preview).

package/config/skills/agent/query-knowledge/skill.json

@@ -0,0 +1,20 @@
+{
+  "id": "agent-query-knowledge",
+  "name": "Query Knowledge",
+  "description": "Search the company knowledge base for SOPs, runbooks, architecture docs, and other documentation.",
+  "category": "memory",
+  "skillType": "claude-skill",
+  "promptFile": "instructions.md",
+  "execution": {
+    "type": "script",
+    "script": {
+      "file": "execute.sh",
+      "interpreter": "bash",
+      "timeoutMs": 15000
+    }
+  },
+  "assignableRoles": ["developer", "qa", "tpm", "designer", "frontend-developer", "backend-developer", "fullstack-dev", "qa-engineer", "product-manager", "architect", "generalist", "sales", "support"],
+  "triggers": ["query knowledge", "search knowledge", "find docs", "search docs", "find runbook", "find SOP"],
+  "tags": ["memory", "knowledge", "search", "documentation"],
+  "version": "1.0.0"
+}

package/config/skills/agent/read-task/execute.sh

@@ -0,0 +1,15 @@
+#!/bin/bash
+# Read the full details of a task file by its absolute path
+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 '{\"absoluteTaskPath\":\"/path/to/task\"}'"
+
+ABSOLUTE_TASK_PATH=$(echo "$INPUT" | jq -r '.absoluteTaskPath // empty')
+require_param "absoluteTaskPath" "$ABSOLUTE_TASK_PATH"
+
+BODY=$(jq -n --arg absoluteTaskPath "$ABSOLUTE_TASK_PATH" '{absoluteTaskPath: $absoluteTaskPath}')
+
+api_call POST "/task-management/read-task" "$BODY"

package/config/skills/agent/read-task/instructions.md

@@ -0,0 +1,19 @@
+# Read Task
+
+Read the full details of a task file by its absolute path. Returns the task content, metadata, status, and any subtasks.
+
+## Parameters
+
+| Parameter | Required | Description |
+|-----------|----------|-------------|
+| `absoluteTaskPath` | Yes | Absolute filesystem path to the task file |
+
+## Example
+
+```bash
+bash config/skills/agent/read-task/execute.sh '{"absoluteTaskPath":"/projects/app/tasks/implement-login.md"}'
+```
+
+## Output
+
+JSON with full task details including description, acceptance criteria, priority, status, and related files.

package/config/skills/agent/read-task/skill.json

@@ -0,0 +1,20 @@
+{
+  "id": "agent-read-task",
+  "name": "Read Task",
+  "description": "Read the full details of a task file by its absolute path.",
+  "category": "task-management",
+  "skillType": "claude-skill",
+  "promptFile": "instructions.md",
+  "execution": {
+    "type": "script",
+    "script": {
+      "file": "execute.sh",
+      "interpreter": "bash",
+      "timeoutMs": 15000
+    }
+  },
+  "assignableRoles": ["developer", "qa", "tpm", "designer", "frontend-developer", "backend-developer", "fullstack-dev", "qa-engineer", "product-manager", "architect", "generalist", "sales", "support"],
+  "triggers": ["read task", "view task", "show task", "task details"],
+  "tags": ["task", "read", "details"],
+  "version": "1.0.0"
+}

package/config/skills/agent/recall/execute.sh

@@ -0,0 +1,24 @@
+#!/bin/bash
+# Retrieve stored memories relevant to a given context
+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 '{\"agentId\":\"dev-1\",\"context\":\"authentication patterns\"}'"
+
+AGENT_ID=$(echo "$INPUT" | jq -r '.agentId // empty')
+CONTEXT=$(echo "$INPUT" | jq -r '.context // empty')
+require_param "agentId" "$AGENT_ID"
+require_param "context" "$CONTEXT"
+
+# Build body with required and optional fields
+BODY=$(echo "$INPUT" | jq '{
+  agentId: .agentId,
+  context: .context
+} +
+  (if .scope then {scope: .scope} else {} end) +
+  (if .limit then {limit: (.limit | tonumber)} else {} end) +
+  (if .projectPath then {projectPath: .projectPath} else {} end)')
+
+api_call POST "/memory/recall" "$BODY"

package/config/skills/agent/recall/instructions.md

@@ -0,0 +1,23 @@
+# Recall
+
+Retrieve stored memories relevant to a given context or query. Use this to look up past decisions, architectural patterns, or findings before starting related work.
+
+## Parameters
+
+| Parameter | Required | Description |
+|-----------|----------|-------------|
+| `agentId` | Yes | Your agent ID |
+| `context` | Yes | Search context or query describing what you want to recall |
+| `scope` | No | Filter by scope: `"project"`, `"team"`, or `"global"` |
+| `limit` | No | Maximum number of results to return |
+| `projectPath` | No | Filter by project path |
+
+## Example
+
+```bash
+bash config/skills/agent/recall/execute.sh '{"agentId":"dev-1","context":"authentication implementation patterns","scope":"project","limit":5}'
+```
+
+## Output
+
+JSON array of matching memory entries with content, category, scope, and timestamps.

package/config/skills/agent/recall/skill.json

@@ -0,0 +1,20 @@
+{
+  "id": "agent-recall",
+  "name": "Recall",
+  "description": "Retrieve stored memories relevant to a given context or query.",
+  "category": "memory",
+  "skillType": "claude-skill",
+  "promptFile": "instructions.md",
+  "execution": {
+    "type": "script",
+    "script": {
+      "file": "execute.sh",
+      "interpreter": "bash",
+      "timeoutMs": 15000
+    }
+  },
+  "assignableRoles": ["developer", "qa", "tpm", "designer", "frontend-developer", "backend-developer", "fullstack-dev", "qa-engineer", "product-manager", "architect", "generalist", "sales", "support"],
+  "triggers": ["recall", "search memory", "remember what", "look up"],
+  "tags": ["memory", "recall", "search", "context"],
+  "version": "1.0.0"
+}

package/config/skills/agent/record-learning/execute.sh

@@ -0,0 +1,29 @@
+#!/bin/bash
+# Record a learning or insight for team knowledge sharing
+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 '{\"agentId\":\"dev-1\",\"agentRole\":\"developer\",\"projectPath\":\"/projects/app\",\"learning\":\"Jest mock resets are needed between tests\"}'"
+
+AGENT_ID=$(echo "$INPUT" | jq -r '.agentId // empty')
+AGENT_ROLE=$(echo "$INPUT" | jq -r '.agentRole // empty')
+PROJECT_PATH=$(echo "$INPUT" | jq -r '.projectPath // empty')
+LEARNING=$(echo "$INPUT" | jq -r '.learning // empty')
+require_param "agentId" "$AGENT_ID"
+require_param "agentRole" "$AGENT_ROLE"
+require_param "projectPath" "$PROJECT_PATH"
+require_param "learning" "$LEARNING"
+
+# Build body with required and optional fields
+BODY=$(echo "$INPUT" | jq '{
+  agentId: .agentId,
+  agentRole: .agentRole,
+  projectPath: .projectPath,
+  learning: .learning
+} +
+  (if .relatedTask then {relatedTask: .relatedTask} else {} end) +
+  (if .relatedFiles then {relatedFiles: .relatedFiles} else {} end)')
+
+api_call POST "/memory/record-learning" "$BODY"

package/config/skills/agent/record-learning/instructions.md

@@ -0,0 +1,24 @@
+# Record Learning
+
+Record a learning or insight gained during task execution. These learnings are shared with the team and accumulated over time to improve future work.
+
+## Parameters
+
+| Parameter | Required | Description |
+|-----------|----------|-------------|
+| `agentId` | Yes | Your agent ID |
+| `agentRole` | Yes | Your role (e.g., `"developer"`, `"qa"`) |
+| `projectPath` | Yes | Absolute path to the project |
+| `learning` | Yes | Description of the learning or insight |
+| `relatedTask` | No | Path or ID of the task that triggered this learning |
+| `relatedFiles` | No | Array of file paths related to the learning |
+
+## Example
+
+```bash
+bash config/skills/agent/record-learning/execute.sh '{"agentId":"dev-1","agentRole":"developer","projectPath":"/projects/app","learning":"Jest mock resets are required between tests when using shared module mocks","relatedTask":"implement-auth-tests","relatedFiles":["src/auth.service.test.ts"]}'
+```
+
+## Output
+
+JSON confirmation that the learning was recorded.

package/config/skills/agent/record-learning/skill.json

@@ -0,0 +1,20 @@
+{
+  "id": "agent-record-learning",
+  "name": "Record Learning",
+  "description": "Record a learning or insight gained during task execution for team knowledge sharing.",
+  "category": "memory",
+  "skillType": "claude-skill",
+  "promptFile": "instructions.md",
+  "execution": {
+    "type": "script",
+    "script": {
+      "file": "execute.sh",
+      "interpreter": "bash",
+      "timeoutMs": 15000
+    }
+  },
+  "assignableRoles": ["developer", "qa", "tpm", "designer", "frontend-developer", "backend-developer", "fullstack-dev", "qa-engineer", "product-manager", "architect", "generalist", "sales", "support"],
+  "triggers": ["record learning", "learned that", "new insight", "knowledge share"],
+  "tags": ["memory", "learning", "knowledge", "insight"],
+  "version": "1.0.0"
+}

package/config/skills/agent/register-self/execute.sh

@@ -0,0 +1,28 @@
+#!/bin/bash
+# Register an agent with the Crewly backend
+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 '{\"role\":\"developer\",\"sessionName\":\"dev-1\"}'"
+
+ROLE=$(echo "$INPUT" | jq -r '.role // empty')
+SESSION_NAME=$(echo "$INPUT" | jq -r '.sessionName // empty')
+require_param "role" "$ROLE"
+require_param "sessionName" "$SESSION_NAME"
+
+# Optional: claudeSessionId for resume support
+CLAUDE_SESSION_ID=$(echo "$INPUT" | jq -r '.claudeSessionId // empty')
+TEAM_MEMBER_ID=$(echo "$INPUT" | jq -r '.teamMemberId // empty')
+
+BODY=$(jq -n \
+  --arg role "$ROLE" \
+  --arg sessionName "$SESSION_NAME" \
+  --arg claudeSessionId "$CLAUDE_SESSION_ID" \
+  --arg teamMemberId "$TEAM_MEMBER_ID" \
+  '{role: $role, sessionName: $sessionName} +
+   (if $claudeSessionId != "" then {claudeSessionId: $claudeSessionId} else {} end) +
+   (if $teamMemberId != "" then {teamMemberId: $teamMemberId} else {} end)')
+
+api_call POST "/teams/members/register" "$BODY"

package/config/skills/agent/register-self/instructions.md

@@ -0,0 +1,18 @@
+# Register Self
+
+Register this agent as active with the Crewly backend. This must be the first skill you run on startup.
+
+## Parameters
+
+| Parameter | Required | Description |
+|-----------|----------|-------------|
+| `role` | Yes | Agent role (e.g., "developer", "qa", "tpm") |
+| `sessionName` | Yes | Your session name (from your identity) |
+| `teamMemberId` | No | Your team member ID |
+| `claudeSessionId` | No | Claude session ID for resume support |
+
+## Example
+
+```bash
+bash config/skills/agent/register-self/execute.sh '{"role":"developer","sessionName":"dev-1"}'
+```

package/config/skills/agent/register-self/skill.json

@@ -0,0 +1,20 @@
+{
+  "id": "agent-register-self",
+  "name": "Register Self",
+  "description": "Register the agent as active with the Crewly backend. Must be called on startup.",
+  "category": "system",
+  "skillType": "claude-skill",
+  "promptFile": "instructions.md",
+  "execution": {
+    "type": "script",
+    "script": {
+      "file": "execute.sh",
+      "interpreter": "bash",
+      "timeoutMs": 15000
+    }
+  },
+  "assignableRoles": ["developer", "qa", "tpm", "designer", "frontend-developer", "backend-developer", "fullstack-dev", "qa-engineer", "product-manager", "architect", "generalist", "sales", "support"],
+  "triggers": ["register", "check in", "go online"],
+  "tags": ["system", "registration", "startup"],
+  "version": "1.0.0"
+}

package/config/skills/agent/remember/execute.sh

@@ -0,0 +1,29 @@
+#!/bin/bash
+# Store a memory entry for future recall
+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 '{\"agentId\":\"dev-1\",\"content\":\"The auth module uses JWT\",\"category\":\"architecture\",\"scope\":\"project\"}'"
+
+AGENT_ID=$(echo "$INPUT" | jq -r '.agentId // empty')
+CONTENT=$(echo "$INPUT" | jq -r '.content // empty')
+CATEGORY=$(echo "$INPUT" | jq -r '.category // empty')
+SCOPE=$(echo "$INPUT" | jq -r '.scope // empty')
+require_param "agentId" "$AGENT_ID"
+require_param "content" "$CONTENT"
+require_param "category" "$CATEGORY"
+require_param "scope" "$SCOPE"
+
+# Build body with required and optional fields
+BODY=$(echo "$INPUT" | jq '{
+  agentId: .agentId,
+  content: .content,
+  category: .category,
+  scope: .scope
+} +
+  (if .projectPath then {projectPath: .projectPath} else {} end) +
+  (if .metadata then {metadata: .metadata} else {} end)')
+
+api_call POST "/memory/remember" "$BODY"

package/config/skills/agent/remember/instructions.md

@@ -0,0 +1,24 @@
+# Remember
+
+Store a memory entry for future recall. Use this to persist important context, decisions, architectural findings, or patterns you discover during work.
+
+## Parameters
+
+| Parameter | Required | Description |
+|-----------|----------|-------------|
+| `agentId` | Yes | Your agent ID |
+| `content` | Yes | The content to remember |
+| `category` | Yes | Memory category (e.g., `"architecture"`, `"decision"`, `"pattern"`, `"bug"`) |
+| `scope` | Yes | Memory scope: `"project"`, `"team"`, or `"global"` |
+| `projectPath` | No | Associated project path |
+| `metadata` | No | Additional metadata object |
+
+## Example
+
+```bash
+bash config/skills/agent/remember/execute.sh '{"agentId":"dev-1","content":"Auth module uses JWT with 24h expiry. Refresh tokens stored in httpOnly cookies.","category":"architecture","scope":"project","projectPath":"/projects/app"}'
+```
+
+## Output
+
+JSON confirmation that the memory entry was stored.