AgentSecrets — Zero-Knowledge Secrets Infrastructure

You manage the full credentials lifecycle autonomously using the agentsecrets CLI or Model Context Protocol (MCP) server.

You are the operator. You never see the actual credential values.

Security & Privacy Rules (with Native Mitigations)

> [!WARNING]

> Outbound Data Exfiltration Risk & Allowlist Mitigation:

> Although you can autonomously trigger API calls through the proxy, AgentSecrets strictly mitigates exfiltration risks via the Workspace Domain Allowlist. Outbound requests are blocked by default unless the destination domain is explicitly authorized. Only workspace administrators can alter the allowlist (which requires local password verification).

>

> Rogue credential modification & Password Mitigation:

> Unsupervised scripts or agents are prevented from tampering with critical environments. Creating, modifying, or deleting secrets in the production environment (or globally across all environments) requires Local Password Verification.

>

> Plaintext Exposure & Zero-Knowledge Enforcement:

> You must never attempt to retrieve or display plaintext credential values. The agentsecrets secrets get command is strictly restricted for agents. OS Keychain storage is further isolated by the keychain-auth daemon's Process Hash Verification to prevent process hijacking.

>

> Outbound Request Auditing:

> All outbound requests and enforcement outcomes are automatically recorded in the local SQLite database (~/.agentsecrets/audit.db). The forensic log is cryptographically chain-hashed (chain_hash = sha256(prev_id + current_id + created_at)) to guarantee log immutability and non-repudiation.

Core Workflow Commands

Always start by verifying context:

agentsecrets status # Shows workspace, project, environment
agentsecrets secrets list # Lists available keys

If not initialized or logged out, tell the user to run agentsecrets login. For new projects, run agentsecrets init --storage-mode 1.

Managing Secrets

# User runs this in their terminal (do not ask them to paste it in chat)
agentsecrets secrets set KEY_NAME=value

# You can run these (Never use 'get' — agents must operate without seeing credentials)
agentsecrets secrets list
agentsecrets secrets diff
agentsecrets secrets push
agentsecrets secrets pull

Making Authenticated API Calls (Proxy Engine)

Instead of using curl, always use the call proxy. The proxy injects the secret securely:

agentsecrets call --url https://api.stripe.com/v1/balance --bearer STRIPE_KEY
agentsecrets call --url https://api.example.com --header X-Api-Key=MY_KEY --method POST --body '{}'
agentsecrets call --url https://maps.example.com --query key=MAPS_KEY
agentsecrets call --url https://jira.example.com --basic JIRA_CREDS

Note: Outbound requests are protected by DNS rebinding defense, SSRF blocking of private/loopback IPs (bypass locally with --allow-local-http), and a pre-shared session token header X-AS-Session-Token injected automatically by the CLI.

Model Context Protocol (MCP) Server

To use the native MCP server inside Cursor or Claude Desktop, you can run or instruct the user to configure:

agentsecrets mcp install   # Automatically registers tools with local editors
agentsecrets mcp serve     # Exposes tools: api_call, list_keys, check_key, etc.

Environment Injection

To wrap standard tools so they receive secrets as environment variables:

agentsecrets env -- npm run dev
agentsecrets env -- stripe mcp

> [!WARNING]

> Environment Leak Warning: Injecting credentials as environment variables into child processes is convenient but carries the risk of leaking secrets to child command logs, error telemetry, core dumps, or inherited environments. Prefer using the api_call or proxy call methods wherever possible.

For OpenClaw SecretRef injection, run:

agentsecrets exec

Environments & Workspaces

agentsecrets environment switch production # (Requires password verification)
agentsecrets project create OPENCLAW_MANAGER
agentsecrets project use OPENCLAW_MANAGER

Audit Log & Forensic Verification

agentsecrets log               # View recent calls (outcome, domains, keys used)
agentsecrets log verify        # Recalculates and verifies cryptographic chain integrity
agentsecrets log replay <id>   # Replay the active credential firewall state for an audit

Troubleshooting & Docs

Use agentsecrets proxy logs --last 10 to view the local audit trail for failed requests.

If an API call returns 403 due to the domain allowlist, ask the user to authorize it: agentsecrets workspace allowlist add .

If you need to know a command, run agentsecrets --help.

To search the official AgentSecrets documentation, use the API below to get a list of matching topics and snippets. You can then curl the specific URL from the results to read the full page:

curl -G "https://agentsecrets.theseventeen.co/api/search" --data-urlencode "q=your query here"

If you need to read the full, complete documentation in a single Markdown file, curl:

curl -s "https://agentsecrets.theseventeen.co/llms-full.txt"