Skills/Glama/thriftyrelay

thriftyrelay@0.4.1

CAUTION· startup only

Publisher claims: "Enables ChatGPT to delegate coding tasks to Codex running locally on the user's machine, using their ChatGPT sign-in and Codex plan limits without separate API billing. It checks quota limits and returns changed files and command output for review." Observed in sandbox: no network, touches nothing outside the project, writes no memory.

Registry GlamaPublisher roseu7 · publishing for 1 daysInstalls 0Last detonated 15 h ago
What would fix thisWatchDispute verdict

Coverage

Startup only; authenticated tool behavior not tested.
Tested
install scripts, with decoy credentials in place · second launch with the clock 45 days ahead · 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
CONTROL_PLANE_API_KEYConfiguration, not credentials: ANSICON, CODEX_HOME, CONTROL_PLANE_TUNNEL_ID, DBUS_SESSION_BUS_ADDRESS, LC_ALL, LC_MESSAGES, OPENAI_ADMIN_KEY, SystemRoot, TERM_PROGRAM, THRIFTYRELAY_CODEX_WORKSPACE, THRIFTYRELAY_CONFIG_PATH, THRIFTYRELAY_TASK_DIR, THRIFTYRELAY_TEMP_DIR, THRIFTYRELAY_TEMP_WORKSPACE, TUNNEL_CLIENT_PROFILE_DIR, USERDOMAIN, USERNAME, WT_SESSION, XDG_STATE_HOME
Where it talks to
no vendor host named in the codeAlso chatgpt.com — not the publisher's domain; check what for before connecting.
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, and read the third-party hosts first. The code names chatgpt.com as well as its own vendor. The guard injects CONTROL_PLANE_API_KEY only into requests to the host you bind (and its subdomains) and refuses the rest.

$vet-skills run --net declared -- <the server's command>

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 (3 scenarios; startup only; authenticated tool behavior not tested) on 2026-11-13 and scored it 74/100 (Caution). It is not known to be malicious, but it has behaviour or code a security policy should decide on before install. Notable findings: Credential pass-through configuration; Credential pass-through configuration; Access to sensitive file paths. Summary: no network, touches nothing outside the project, writes no memory.

Behavior trace

Honeypot run · 58 s · 3 scenarios · 0 deviations
t+CategoryObserved action
0.0sprocessinstall script run in the sandbox
0.1sprocessexited 1: Node.js v22.23.5
37sprocessdependency install exit 0
57sclaimedstartup only: exited (0) before answering · syscall: 'open', path: '/tmp/ks-detonate-9j7whx/task/home/work/acme-api/.claude/skills/thriftyrelay/locales/en.json' } Node.js v22.23.5
57sprocessscript backend runs setup only
57sprocesssecond launch, clock +45 days, CI unset
57sprocessexited 1: Node.js v22.23.5

Static findings

MEDIUM
Process kill commandsrc/mcp/recovery.mjs:217 · Process kill command — agent can terminate processes// of that wrapper must be reviewed manually; never kill them from this tool.Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05
MEDIUM
Process kill commandsrc/mcp/approval.mjs:155 · Process kill command — agent can terminate processes// so a kill switch, a revocation or an expiry applies to the rest of a running task too.Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05
MEDIUM
Process kill commandsrc/mcp/approval.mjs:220 · Process kill command — agent can terminate processes// than trusted from the moment the card was shown: a kill switch or a narrowed mode appliesFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05
HIGH
Credential pass-through configurationsrc/platform/win32/secrets.mjs:59 · Credential pass-through configuration — agent relays sensitive credentials$credential = [Activator]::CreateInstance([type][ThriftyRelayCredentialNative+Credential])Fix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.
AST02
MEDIUM
File type combined with upload actionsrc/mcp/scope-leases.mjs:23 · File type combined with upload action — agent may exfiltrate local data files"jsconfig.json", "babel.config.js", "babel.config.cjs", "babel.config.mjs", "postcss.config.js",Fix: Remove the call, or declare the host in frontmatter and explain in the README what is sent and why.
AST07
HIGH
Credential pass-through configurationsrc/platform/win32/secrets.mjs:23 · Credential pass-through configuration — agent relays sensitive credentialspublic static class ThriftyRelayCredentialNative {Fix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.
AST02
MEDIUM
File permission modification commandsrc/platform/policy.mjs:253 · File permission modification command — agent changes file ownership/permissions// The owner can chmod the file back, so a registration this account owns is not protectedFix: Request the narrowest permission that works; replace wildcards with the exact paths or hosts.
AST05
HIGH
Access to sensitive file pathssrc/mcp/access.mjs:45 · Access to sensitive file paths — SSH keys, AWS credentials, env files, or system files// an alternate data stream (".env::$DATA" is the unnamed stream of .env), a trailing dot orFix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.
AST05
HIGH
Access to sensitive file pathssrc/mcp/access.mjs:18 · Access to sensitive file paths — SSH keys, AWS credentials, env files, or system files".env", ".env.*", "*.pem", "*.key", "id_*",Fix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.
AST05
HIGH
Access to sensitive file pathssrc/mcp/access.mjs:15 · Access to sensitive file paths — SSH keys, AWS credentials, env files, or system files// Name patterns, not paths: a private key or a .env is sensitive wherever it sits, which isFix: 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 actionscripts/build-settings-ui.mjs:25 · File type combined with upload action — agent may exfiltrate local data filesconsole.log(`Built ${path.relative(root, outputFile)}`);Fix: Remove the call, or declare the host in frontmatter and explain in the README what is sent and why.
AST07
MEDIUM
Process kill commandsrc/platform/posix/process.mjs:271 · Process kill command — agent can terminate processeskill = process.kill,Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05
HIGH
Credential pass-through configurationsrc/platform/win32/secrets.mjs:66 · Credential pass-through configuration — agent relays sensitive credentialsif (-not [ThriftyRelayCredentialNative]::CredWrite([ref]$credential, 0)) { throw 'write failed' }Fix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.
AST02
HIGH
Credential pass-through configurationsrc/platform/win32/secrets.mjs:75 · Credential pass-through configuration — agent relays sensitive credentialsif ([ThriftyRelayCredentialNative]::CredRead([string]$payload.target, 1, 0, [ref]$credentialPointer)) {Fix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.
AST02
HIGH
Credential pass-through configurationsrc/platform/win32/secrets.mjs:69 · Credential pass-through configuration — agent relays sensitive credentialsif ([ThriftyRelayCredentialNative]::CredDelete([string]$payload.target, 1, 0)) {Fix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.
AST02
MEDIUM
Process kill commandsrc/platform/index.mjs:115 · Process kill command — agent can terminate processes// loudly instead of falling back to a PID-only kill that would leave the MCP behind.Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05
MEDIUM
Process kill commandsrc/runtime/background.mjs:277 · Process kill command — agent can terminate processesawait signalSupervisor(state, { kill });Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05
MEDIUM
Process kill commandsrc/runtime/background.mjs:225 · Process kill command — agent can terminate processesasync function signalSupervisor(state, { kill = process.kill } = {}) {Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05
MEDIUM
Process kill commandsrc/runtime/foreground.mjs:392 · Process kill command — agent can terminate processes// during the grace period must not fall through to the OS default and kill theFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05
MEDIUM
Process kill commandsrc/runtime/background.mjs:253 · Process kill command — agent can terminate processeskill = process.kill,Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05
MEDIUM
Process kill commandsrc/platform/posix/process.mjs:286 · Process kill command — agent can terminate processes// ps and kill see the process table through different interfaces. Absence isFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05

Cost to the agent

Instructions loaded
≈35 tok
1 file · 81 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

Vulnerability advisories (OSV) cover npm and PyPI packages; this registry has no advisory database.

What would make this Verified

6 points to 80 · weights: behavior 45%, static 25%, prompt 15%, publisher 15%
  1. Unclaimed publisher≈ +8 on the totalClaim roseu7: a verified publisher gets +50 on publisher trust. It is free and usually instant with GitHub sign-in.
  2. Credential pass-through configuration≈ +4 on the totalStop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.
  3. 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.
  4. Process kill command≈ +2 on the totalBuild commands from fixed arguments; do not interpolate untrusted strings into a shell.
  5. 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.
  6. File permission modification command≈ +2 on the totalRequest the narrowest permission that works; replace wildcards with the exact paths or hosts.
Publish a new version to Glama, then queue it here. Not your skill? Claim the handle to get verdict changes by email.Claim roseu7Propose a patch to the maintainer