Graphiti

Knowledge graph operations via Graphiti API. Search facts, add episodes, and extract entities/relationships.

Security Scan

The skill's files and runtime instructions are consistent with its stated purpose (calling a Graphiti REST API to search/add facts) and do not request unrelated credentials or installs.

Purpose & Capability

Name/description (knowledge-graph operations) match the files and SKILL.md: the skill runs curl against a Graphiti service and offers search/add endpoints. Declared dependencies (graphiti, neo4j, qdrant) align with the stated purpose.

Instruction Scope

Runtime instructions perform network calls to the configured Graphiti URL (/facts/search and /messages) and use a small env-check.sh to discover the URL via clawdbot config or $GRAPHITI_URL. This is expected for a remote API integration, but it means user queries and added 'episodes' will be transmitted to the configured Graphiti service — verify that endpoint is trusted.

Install Mechanism

Instruction-only skill with no install spec; no arbitrary downloads or archive extraction. Manifest lists expected system utilities (curl, jq) but the skill does not install any code automatically.

Credentials

The skill does not request secrets or extra environment variables. It reads a single clawdbot config key (skills.graphiti.baseUrl) and $GRAPHITI_URL which is proportional to its purpose. No unrelated credentials or config paths are requested.

Persistence & Privilege

always is false and the skill does not request elevated or persistent privileges or modify other skills/settings. The env-check script only reads a single config key and checks service health.

Assessment

This skill is coherent for calling a Graphiti API, but be aware it will send whatever you query or add to the configured Graphiti service. Before installing, confirm the Graphiti URL (clawdbot config or $GRAPHITI_URL) points to a trusted server (especially if remote). Ensure curl and jq are available if you intend to run the provided commands. If you don't want your data sent off-host, do not configure a remote Graphiti URL or avoid using the skill.

Graphiti Knowledge Graph

Query and manage your knowledge graph using Graphiti's REST API with dynamic service discovery.

Prerequisites

• Neo4j database (graph storage)
• Qdrant (vector search)
• Graphiti service running (default: http://localhost:8001)

Tools

graphiti_search

Search the knowledge graph for relevant facts.

Usage:

bash command:"
GRAPHITI_URL=\$({baseDir}/references/env-check.sh)
curl -s -X POST \"\$GRAPHITI_URL/facts/search\" \
  -H 'Content-Type: application/json' \
  -d '{\"query\": \"YOUR_QUERY\", \"max_facts\": 10}' | jq .
"

graphiti_add

Add a new episode/memory to the knowledge graph.

Usage:

bash command:"
GRAPHITI_URL=\$({baseDir}/references/env-check.sh)
curl -s -X POST \"\$GRAPHITI_URL/messages\" \
  -H 'Content-Type: application/json' \
  -d '{\"name\": \"EPISODE_NAME\", \"content\": \"EPISODE_CONTENT\"}' | jq .
"

Dynamic Configuration

The skill uses environment discovery to find Graphiti automatically:

1. Clawdbot config: clawdbot config get skills.graphiti.baseUrl
2. Environment variable: $GRAPHITI_URL
3. Default fallback: http://localhost:8001

To change the Graphiti URL:

export GRAPHITI_URL="http://10.0.0.10:8001"
# OR
clawdbot config set skills.graphiti.baseUrl "http://10.0.0.10:8001"

Examples

Search for information:

bash command:"
GRAPHITI_URL=\$({baseDir}/references/env-check.sh)
curl -s -X POST \"\$GRAPHITI_URL/facts/search\" \
  -H 'Content-Type: application/json' \
  -d '{\"query\": \"Tell me about Essam Masoudy\", \"max_facts\": 5}'
"

Add a memory:

bash command:"
GRAPHITI_URL=\$({baseDir}/references/env-check.sh)
curl -s -X POST \"\$GRAPHITI_URL/messages\" \
  -H 'Content-Type: application/json' \
  -d '{\"name\": \"Project Update\", \"content\": \"Completed Phase 1 of Clawdbot integration\"}'
"