Skills/npm/agy-mcp-bridge

agy-mcp-bridge@0.1.38

CAUTION· startup only

Publisher claims: "Local loopback HTTP MCP bridge that delegates bounded coding tasks to multiple agent CLIs (Antigravity, Codex) with optional parallel git worktrees" Observed in sandbox: no network, touches nothing outside the project, writes no memory.

Registry npm (MCP server)Publisher thanhdo257205 · publishing for 24 daysInstalls 7,393Last detonated 3d ago
What would fix thisWatchDispute verdict

Coverage

Startup only; authenticated tool behavior not tested.
Tested
server startup over stdio (initialize sent) · every file read, file written and host contacted by what ran
Not tested
the tool list (the server exited before answering) · what the tools do when called with real credentials

What this means for the label: Caution here covers what ran. The server's tools were not exercised with a real credential, so nothing about their behavior is asserted; the credential summary below says what to do with that.

Before you give it your token

It will ask for
MAX_BUDGET_TOKENSMIN_BUDGET_TOKENSConfiguration, not credentials: ANTIGRAVITY_BRIDGE_CONFIG, BRIDGE_AGY_PERMISSIONS, BRIDGE_DELEGATION, BRIDGE_HOME, BRIDGE_MCP_URL, BRIDGE_NO_SKILL_PROMPT, BRIDGE_NO_UPDATE_CHECK, BRIDGE_OPEN_BROWSER, BRIDGE_UPDATE_CHECK, BRIDGE_UPDATE_REGISTRY, CODEX_HOME, PATHEXT, SERVER_KEY, SystemRoot, XERJ_DEFAULT_INDEX, XERJ_URL
Where it talks to
no vendor host named in the code
What the honeypot saw
no credential files read · no network
What we could not test
the tool list (the server exited before answering) · what the tools do when called with real credentials

Connect it with a scoped credential, under the guard. It asks for MAX_BUDGET_TOKENS, MIN_BUDGET_TOKENS and talks to its vendor. Give the credential to the guard, not the server: the value never enters the server's environment and only travels to the host you bind. What the tools do with a real credential was not observed.

$vet-skills run --net declared -- npx -y agy-mcp-bridge

Never paste a credential into a chat, a ticket or a report, including ours. How the guard holds it.

Is [email protected] safe to install?

KillerSkills ran it in the KillerSkills honeypot sandbox (2 scenarios; startup only; authenticated tool behavior not tested) on 2026-11-11 and scored it 79/100 (Caution). It is not known to be malicious, but it has behaviour or code a security policy should decide on before install. Notable findings: Access to sensitive file paths; Access to sensitive file paths; Access to sensitive file paths. Summary: no network, touches nothing outside the project, writes no memory.

Behavior trace

Honeypot run · 24 s · 2 scenarios · 0 deviations
t+CategoryObserved action
0.0sprocessno install scripts in bundle
3.4sprocessdependency install exit 0
23sclaimedstartup only: exited (0) before answering · [bridge] Health check: http://127.0.0.1:3000/health [bridge] No authentication: any local process running as this user can control the bridge. Keep it on this machine. [bridge] Web UI: http://127.0.0.1:3000/ui (open in a browser on this machine) [bridge] Client disconnected: tasks continue. Ctrl+C: cancel owned tasks and stop server.
23sprocessscript backend runs setup only
23sprocessno install scripts to re-run

Static findings

HIGH
Access to sensitive file pathsdist/src/agents/opencode.js:40 · Access to sensitive file paths — SSH keys, AWS credentials, env files, or system filesreturn { ...env };Fix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.
AST05
MEDIUM
npx invocation of non-well-known packagedist/src/cli.js:71 · npx invocation of non-well-known package — may execute untrusted codeconsole.error(`${notice} This is not a global npm install; update it the way you installed it (for example: npm install -g ${pkg.name}@latest, or npx ${pkg.name}@latest).`);Fix: Pin versions and sources; vendor anything fetched at install time.
AST06
HIGH
Access to sensitive file pathsdist/src/antigravity/process.js:93 · Access to sensitive file paths — SSH keys, AWS credentials, env files, or system filesreturn { ...env, PATHEXT: '.COM;.EXE;.BAT;.CMD' };Fix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.
AST05
MEDIUM
File type combined with upload actiondist/scripts/probe-stream.js:54 · File type combined with upload action — agent may exfiltrate local data filesconsole.log(`agy exit code: ${code}; ${lines.length} stdout line(s); raw output saved to ${path.join(results, 'stream-probe.ndjson')} (not committed)`);Fix: Remove the call, or declare the host in frontmatter and explain in the README what is sent and why.
AST07
HIGH
Access to sensitive file pathsdist/src/agents/opencode.js:45 · Access to sensitive file paths — SSH keys, AWS credentials, env files, or system filesreturn { ...env, OPENCODE_CONFIG_CONTENT: JSON.stringify({ ...config, permission: { ...current, [permission]: 'allow' } }) };Fix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.
AST05
MEDIUM
Process kill commandskills/antigravity-bridge/SKILL.md:128 · Process kill command — agent can terminate processes…On a lock error, reconcile existing tasks too. Do not remove execution locks or kill processes by name. If the lock looks stale, tell the user that the bridge owner …Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05
MEDIUM
Process kill commanddist/src/isolation/orphans.js:240 · Process kill command — agent can terminate processes// Enumeration and termination are not one atomic act: every kill is awaited, so a pid can be reissued while theFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05
MEDIUM
Process kill commanddist/src/isolation/orphans.js:239 · Process kill command — agent can terminate processesconst kill = options.kill ?? killTree;Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05
MEDIUM
File type combined with upload actiondist/src/setup.js:25 · File type combined with upload action — agent may exfiltrate local data files… ' Freebuff starts the bridge itself over stdio. Merge this into ~/.agents/mcp.json, or into .agents/mcp.json of the project you open in Freebuff (it reads the nearest of the two):', '', ...entry, '', ' That bridge is a child process of Fr…Fix: Remove the call, or declare the host in frontmatter and explain in the README what is sent and why.
AST07
MEDIUM
File type combined with upload actiondist/src/setup.js:21 · File type combined with upload action — agent may exfiltrate local data files…-scope local antigravity ${url}`); } add(' or put this in .mcp.json of the project:', '', ' {', ' "mcpServers": {', ' "antigravity": {', ' "type": "http",', ` "url": "${input.mcpUrl}"`, ' }', ' }', ' }', '', ' Check: claude mcp get an…Fix: Remove the call, or declare the host in frontmatter and explain in the README what is sent and why.
AST07
MEDIUM
npx invocation of non-well-known packagedist/src/cli.js:231 · npx invocation of non-well-known package — may execute untrusted code…{ console.error(`Not changed: this copy of ${BIN} runs from a temporary npx cache (${paths.packageRoot}), which npm can delete, so Freebuff's config would point …Fix: Pin versions and sources; vendor anything fetched at install time.
AST06
MEDIUM
npx invocation of non-well-known packagedist/src/cli.js:221 · npx invocation of non-well-known package — may execute untrusted code…geRoot)) console.error(`Note: this copy of ${BIN} runs from a temporary npx cache, so the stdio command printed below stops working once npm clears it. Install it…Fix: Pin versions and sources; vendor anything fetched at install time.
AST06
MEDIUM
Process kill commandskills/mission-orchestrator/SKILL.md:8 · Process kill command — agent can terminate processes…r the user resumes it. Read `antigravity-bridge` first: delegation consent, the kill switch, task observation, permissions and review rules all apply unchanged to mi…Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05
MEDIUM
Process kill commanddist/src/delegation.js:6 · Process kill command — agent can terminate processes* Owner-controlled kill switch for new delegations, checked on every call so no restart is needed.Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05
MEDIUM
MCP server using network transport (SSE/HTTP)dist/src/setup.js:21 · MCP server using network transport (SSE/HTTP) — data sent over network… add(' or put this in .mcp.json of the project:', '', ' {', ' "mcpServers": {', ' "antigravity": {', ' "type": "http",', ` "url": "${input.mcpUrl}"`, ' }', ' }', ' }', …Fix: Namespace tool names and keep tool descriptions literal; no instructions to the agent inside them.
AST04
MEDIUM
Process kill commandskills/antigravity-bridge/SKILL.md:22 · Process kill command — agent can terminate processes…ser can end it at any time. The cases in "When NOT to delegate" (a decline, the kill switch) still win over it. In this mode, work like this to keep your own token …Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05
MEDIUM
Process kill commanddist/src/tasks/manager.js:1224 · Process kill command — agent can terminate processes// Diagnose only; never remove a lock automatically, never kill anything. PIDs can be reused, so wording stays cautious.Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05
LOW
Fake instruction boundary tagdocs/OPENCODE-INTEGRATION-PLAN.md:15 · Fake instruction boundary tag · informational: not an instruction file| `opencode run <prompt>` qua argv | `opencode run --standalone --format json`, prompt qua **stdin** (argv bị giới hạn độ dài, và prompt bắt đầu bằng `-` bị đọc thành cờ) |Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.
AST01

Cost to the agent

Instructions loaded
≈8,747 tok
2 files · 75 in bundle
Tool list per prompt
≈0 tok
0 tools · 0 per tool
Handshake
no reply
launch to tools/list, stdio
Undocumented tools
0
description under 20 chars · avg 0

Tokens estimated at four characters each. The tool list is re-read by the agent on every turn, so it is the number to compare between servers that do the same job: ranked across the index.

Known vulnerabilities

No advisories in OSV for agy-mcp-bridge@0.1.38 as of 3d ago. This is the package itself, not its dependencies.

What would make this Verified

1 point to 80 · weights: behavior 45%, static 25%, prompt 15%, publisher 15%
  1. Unclaimed publisher≈ +8 on the totalClaim thanhdo257205: a verified publisher gets +50 on publisher trust. It is free and usually instant with GitHub sign-in.
  2. Access to sensitive file paths≈ +4 on the totalStop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.
  3. npx invocation of non-well-known package≈ +2 on the totalPin versions and sources; vendor anything fetched at install time.
  4. File type combined with upload action≈ +2 on the totalRemove the call, or declare the host in frontmatter and explain in the README what is sent and why.
  5. Process kill command≈ +2 on the totalBuild commands from fixed arguments; do not interpolate untrusted strings into a shell.
  6. MCP server using network transport (SSE/HTTP)≈ +2 on the totalNamespace tool names and keep tool descriptions literal; no instructions to the agent inside them.
Publish a new version to npm, then queue it here. Not your skill? Claim the handle to get verdict changes by email.Claim thanhdo257205Propose a patch to the maintainer