Security audit
Security checks for vulnerabilities and agentic risk
This skill is a disclosed connector for storing and retrieving agent knowledge in a Synap pod, with expected remote storage and write authority for that purpose.
Install only if you are comfortable with your agent storing facts, documents, relationships, and relayed channel messages in the configured Synap pod. Use least-privilege API keys where available, review managed-pod privacy expectations, and configure governance so important business data changes require review.
Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
git clone https://github.com/synap-core/backend && cd backend/deploy
cp .env.example .env
docker compose up -d
# Visit https://your-domain/registration to create account
./setup-openclaw.sh
Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).
GET /api/hub/memory?userId={UID}&query=Marc+preferences&limit=10
DELETE /api/hub/memory/?userId={UID}
### Create & Update
Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).
"sourceEntityId": "<id>", "targetEntityId": "<id>", "type": "related",
"reasoning": "Marc works at Acme Corp" }
DELETE /api/hub/relations/<id>
{ "userId": "{UID}", "agentUserId": "{UID}", "workspaceId": "{WID}" }
POST /api/hub/profiles
The README advertises message relay from third-party platforms and remote API-key-based connectivity, but it does not warn users that messages, documents, facts, and other agent data may be transmitted to and stored on a remote Synap pod. In an agent skill context, missing disclosure can lead to unintended exfiltration of sensitive chat content, credentials, or long-term memory to external infrastructure, especially when managed hosting is encouraged.
The skill explicitly instructs the agent to relay third-party messages from external platforms into Synap and trigger Synap AI processing, but it does not require a clear user-facing consent or privacy warning before transmitting that content. This creates a real data-handling risk because messages from Telegram, WhatsApp, or Slack may contain sensitive personal or business information that is forwarded to another system for storage and automated processing.
Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.
2. **Recall** — search across everything by keyword, type, or semantic similarity
3. **Relay** — bring external messages (Telegram, WhatsApp, Slack) into Synap and let the Synap AI process them
Every write goes through governance — some auto-approve, others create proposals
for the user to review. You never lose data. You never need to organize it.
---
Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.
2. **Recall** — search across everything by keyword, type, or semantic similarity
3. **Relay** — bring external messages (Telegram, WhatsApp, Slack) into Synap and let the Synap AI process them
Every write goes through governance — some auto-approve, others create proposals
for the user to review. You never lose data. You never need to organize it.
---
Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.
| Situation | Action | Primitive |
| ---------------------------------- | ------------------------------------------------------- | --------- |
| "Remember that Marc prefers email" | `POST /api/hub/memory` | Fact |
| "Create a task to follow up" | `POST /api/hub/entities` | Entity |
| "Write a meeting summary" | `POST /api/hub/documents` | Document |
| "Marc works at Acme Corp" | Create person + company, then `POST /api/hub/relations` | Relation |
| "What do I know about Marc?" | `GET /api/hub/search?query=Marc` | Search |
Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.
## Governance
### Auto-approved:
- All reads and searches
- Memory storage/recall
Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.
## Governance
### Auto-approved:
- All reads and searches
- Memory storage/recall
No suspicious patterns detected.