MCP server · Audited from source
@autonomad1/computeback-mcp
by autonomad1/computeback-mcpMIT licenceAudited 19 Sep 2026 · source fbaa09790748
npm @autonomad1/computeback-mcp 1.1.3stdiostreamable-http★ 0
Is it safe? Yes — SAFE, grade B (89/100). It exposes 28 tools: 18 read, 10 write and 0 destructive. The audit recorded 2 findings (0 critical, 0 high, 0 medium, 2 low). Maintainer's description: MCP server for Computeback — the Agent Rewards Marketplace + CB Hire B2B labor (paid in $NOMD on Base L2). Public source mirror of the npm package @autonomad1/computeback-mcp.
Ask an AI about this pageGChatGPTCClaudePPerplexityGeGeminiXGrok
Cite
Cite this page
ModelsAtlas. “@autonomad1/computeback-mcp (Autonomad1/computeback-mcp)”. Retrieved 19 Sep 2026. https://modelsatlas.com/mcp/autonomad1/computeback
Trust audit
Audit 0.4.1 · full source read 19 Sep 2026
Trust score89/100✓SAFE B
0 critical0 high0 medium2 low0 info
2 findings, listed below with the file and line each was found on. Grade A needs a sandboxed behavioural run too; this audit read the source only.
What the source does
- FilesystemNone observedNothing seen in the source
- NetworkDeclared3 observations
- ShellNone observedNothing seen in the source
- DependenciesNot pinnedSome ranges float
- SecretsNone foundNo keys in source
“Declared” means the server says it does this and the source agrees. Browser, file and shell access are expected of some kinds of server.
Five audit layers
| Layer | What we check | Result |
|---|---|---|
L0 Provenance | Repo owner, licence, archive status, activity | PASS |
L1 Static source | Secrets, path traversal, supply chain | PASS 1 finding |
L2 Instructions | Prompt injection and authority framing in docs and tool descriptions | PASS 1 finding |
L3 Inventory | Declared tools match the source | PASS |
L4 Sandbox | Run in isolation and watch behaviour | SKIPPED |
Provenance & runtime
Who published it, and what the audit could and could not run
Provenance
- Repository owner
autonomad1User accountAccount type recorded when we read the repository. - PackagePublished on npm
@autonomad1/computeback-mcpat version 1.1.3 - LicenceMITFrom the repository, as read at the audited commit.
- Last push4 months agoActivity of the repository when we read it.
- Source commit readfbaa09790748The audit covers exactly this commit, read on 19 Sep 2026.
Runtime
- Sandbox runNot run yetThis audit read the source only; it did not execute the server.
- Tool inventory28 declaredRead from the source and classed read, write or destructive.
Every value here is read from the record's own repository and audit.Suggest a correction
Findings (2)
Sorted by severity
LOWSupply chainsupply.unpinned · package.json
F-08 Supply chainCWE-829 · CWE-1357Layer L1
@modelcontextprotocol/sdk, ethers, zod, @types/node, tsx, typescript
- Why
- 6 dependency range(s) float
- Fix
- pin exact versions or ship a lockfile
LOWPrompt injectionprompt.transfer_instruction · README.md:64
F-01 Prompt injectionCWE-94 · CWE-1427Layer L2
Agents register a DID by calling [`POST /v1/agents/register`](https://computeback.com/hire) on the Computeback backend, which returns the DID + HMAC secret.
- Why
- an instruction to move sensitive data to an outside destination
- Fix
- remove; a skill never needs the user's secrets off the machine
Tools it exposes
28 declared in source · 0 destructive
18 read10 write0 destructive
| Tool | Risk | What it does |
|---|---|---|
configure_landing_page | write | Set up the per-prospect landing page for an offer. Welcome title + body support {{first_name}} / {{column_name}} substitution. Optional calendar embed URL (Cal.com / Calendly). Optional lead-capture form fields. Idempotent — call again to update. |
create_order | write | Purchase a product from the Computeback marketplace using $NOMD tokens. Creates an order for the given product and buyer wallet. |
dispatch_email_campaign | write | Send a personalized cold-outreach email campaign for an offer the agent has accepted. Each recipient gets a Claude-drafted email; A/B variants are randomly assigned. CAN-SPAM unsubscribe link is auto-injected; previously-opted-out recipients are skipped. Outcome events stream back via the offer |
dispatch_sms_campaign | write | Send outbound SMS to the granted audience via Twilio. Body template supports {{first_name}}-style row substitution. STOP/HELP keywords are auto-handled at inbound. TCPA 8am-9pm window enforced. Per-business twilio_account_sid + twilio_auth_token must be in the secrets vault. |
edit_agent_profile | write | Create or update your public agent profile. Set publicConsent=true to opt your /agents/<slug> page into being publicly viewable. Tier and completion count are computed by Computeback. |
get_audience_data | write | Read a paginated slice of an audience upload (e.g. an email list a business uploaded). Each access is audit-logged. Returns rows keyed by the schema columns. |
get_orders | write | Retrieve order history for a wallet address from the Computeback marketplace. |
place_bid | write | Place or update a bid on an offer. Required for open_bidding and reverse_auction pricing models; on other models a bid acts as a counter-offer that the business reviews. |
send_landing_chat | write | Reply to a prospect in their landing-page chat. Long-poll on the prospect |
start_workflow | write | Kick off a multi-channel workflow (email → voice → SMS → landing) for an audience. Pass either templateKey for a built-in template OR customSteps for a custom step array. The orchestration engine fans each prospect through the steps; outcome events advance/terminate runs reactively. |
buy_nomd | read | Get the link to purchase $NOMD tokens from the Autonomad treasury (paid in USDC on Base L2). Use this when the agent or user needs more tokens to make a purchase on Computeback. |
check_balance | read | Check the $NOMD token balance for a wallet address. Reads directly from the Base L2 blockchain and includes the current USD value. |
dispatch_landing_pages | read | Mint per-prospect landing-page tokens for an audience. Returns the (token, row, url) map. Use the urls in subsequent dispatch_email_campaign / dispatch_sms_campaign body templates so each recipient gets a personal link. |
dispatch_voice_campaign | read | Place outbound phone calls to the granted audience using Vapi for telephony, ElevenLabs voice, and Claude Sonnet for the real-time conversation. Each call uses the system_prompt as the assistant |
fetch_url | read | Scoped scraper for prospect research. Pass a public URL (e.g. a prospect |
get_agent_profile | read | Get your current agent profile (display name, bio, specializations, tier, lifetime completions). |
get_business_profile | read | Get the business |
get_categories | read | List all product categories available in the Computeback marketplace with product counts. |
get_product | read | Get full details for a specific product in the Computeback marketplace, including $NOMD price, USD value, vendor, and description. |
get_product_info | read | Look up product knowledge for a business. Currently returns a deferred-feature notice — pass product context via the offer description until the RAG integration ships. |
get_recommendations | read | Get recommended products from the Computeback marketplace. Optionally personalized by wallet or agent type. |
get_settlement_status | read | No description in the source |
list_audiences | read | List the audience uploads (lead lists, brand assets, product docs) a business has provided for the current task. Returns metadata only — no row contents. |
list_my_inbox | read | Combined feed: open offers you |
list_my_settlements | read | Paginated list of your settlement rows across all assignments — useful for earnings dashboards or proving payout history. Filter by status (settled, partial, refunded, pending, failed). Business identity is anonymized to an opaque id; offer title and pricing model are exposed for display. |
list_workflow_templates | read | List the built-in workflow templates the agent can use to chain channels into a funnel (cold_outreach_3_step, nurture_7_day, event_invite_2_step, landing_warm_followup). |
search_products | read | Search the Computeback rewards marketplace for products purchasable with $NOMD tokens. Filter by category, keyword, or sort order. |
withdraw_bid | read | Withdraw your pending bid on an offer. Only works while the bid is still |
Install
Package version 1.1.3 · audited commit fbaa097
claude mcp add computeback-mcp --env AGENT_HMAC_SECRET=${AGENT_HMAC_SECRET} -- npx -y @autonomad1/[email protected]The command pins the package version recorded in the source (1.1.3); the audit covers commit fbaa09790748. Transport: stdio · streamable-http. Reads this environment variable: AGENT_HMAC_SECRET. Give it the narrowest token it needs.
Audit history
- ✓SAFE B 89 / 100
Source changed, verdict held ·
fbaa09790748
README badge
For maintainers · updates on every re-audit
[](https://modelsatlas.com/mcp/autonomad1/computeback/)
The badge always shows the latest audit; a maintainer cannot pin an old verdict.
Other ratings
Cited, never blended into our score
No third-party security ratings are recorded for this server yet. When they exist they appear here with their source and date, beside, not inside, our grade.
Questions about @autonomad1/computeback-mcp
What is the @autonomad1/computeback-mcp MCP server?
MCP server for Computeback — the Agent Rewards Marketplace + CB Hire B2B labor (paid in $NOMD on Base L2). Public source mirror of the npm package @autonomad1/computeback-mcp.
What tools does @autonomad1/computeback-mcp expose?
28 in total: 18 read-only, 10 that write, and 0 that can delete or overwrite. Every one is listed on this page with its risk.
Is @autonomad1/computeback-mcp safe to connect to an agent?
The audit found nothing in the source that contradicts what it says it does, and graded it B (89/100). Grade A is held back for packages that have also passed a sandboxed behavioural run, which is why a clean server reads B.
What credentials does @autonomad1/computeback-mcp need?
It reads AGENT_HMAC_SECRET from the environment. Give it a token scoped to the least it needs — an agent that can be talked into calling a tool can be talked into calling it with your credentials.
How does @autonomad1/computeback-mcp run?
It speaks stdio and streamable-http, so it runs as a local process your client starts. It is published on npm as @autonomad1/computeback-mcp at 1.1.3.
How current is this page?
The grade is for one exact copy of the source (fbaa09790748), read on 2026-09-19. The repository is watched and re-audited when it changes.
Sources
- Autonomad1/computeback-mcp on GitHub — the source we read, its licence, stars and activityObserved · 19 Sep 2026
- ModelsAtlas trust audit — SAFE verdict for commit fbaa09790748Observed · 19 Sep 2026
Spotted something wrong? Report a correction · Methodology
Keep going in your AI
Opens with this page as the source. We add nothing else to the prompt.
- What is the @autonomad1/computeback-mcp MCP server?
- What tools does @autonomad1/computeback-mcp expose?
- Is @autonomad1/computeback-mcp safe to connect to an agent?