imagegen-swarm@1.4.5
CAUTION· install onlyPublisher claims: "AI image generation for stills and video (text-to-image, text-to-video, image-to-video, and image edits) by pooling leftover quotas from multiple engines into a single batch. Use when the user asks to generate or edit an image, draw, use reference images, make a video, or run text-to-image, text-to-video, or image-to-video, or mentions Codex, Grok, Grok Imagine, Gemini, nano banana, Antigravity, Jimeng, Dreamina, Muse, Lovart, gpt-image, or a pay-per-image API. Enabled engines start together on one queue and keep running if one runs out of quota. Jimeng and Grok can generate video; Jimeng also does image-to-video from the first reference and edits with up to 10 reference images; Antigravity Gemini edits with up to 3. An optional paid OpenAI or Gemini API line runs only when a job specifies api, or when no enabled subscription engine can take that job. For a shared character, brand, mascot, or cover series, ask before enabling series mode. Not for ordinary local crop, resize, or convert tasks." Observed in sandbox: no network, touches nothing outside the project, writes no memory.
Coverage
Bundled scripts ran with no arguments; no agent task scenario for this bundle.- Tested
- 8 of 25 bundled scripts run with no arguments (the rest were not run: per-bundle cap), with decoy credentials in the environment · every file read, file written and host contacted by what ran
- Not tested
- use of the skill by an agent
What changed since 1.4.4
vet-skills diff imagegen-swarm 1.4.4 1.4.5In the sandbox, this version reached for nothing 1.4.4 did not. Compared from the two recorded runs, not from source.
1.4.4 was not run in the sandbox (static only); the comparison covers what static analysis recorded only.
The guard can hold an update like this until someone approves it (runtime.on_change: hold). A change that only shows with a real credential, or on a path the run did not reach, is not in a trace and so not in this comparison.
Is [email protected] safe to install?
KillerSkills ran it in the KillerSkills honeypot sandbox (2 scenarios; bundled scripts ran with no arguments; no agent task scenario for this bundle) on 2026-11-22 and scored it 69/100 (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; Collects environment variables by secret-like names. Summary: no network, touches nothing outside the project, writes no memory.
Behavior trace
Honeypot run · 1 s · 2 scenariosStatic findings
? { ...env, [probe.keyEnv]: process.env[probe.keyEnv] } : env;Fix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.…ck | Same as the engine it probes. For Codex it only tests that `CODEX_HOME/auth.json` exists | No prompts or images in probe-only mode. The paid-line check sends only the key | | `smoke.mjs` or `check.mjs --live` | The checks above and th…Fix: Remove the call, or declare the host in frontmatter and explain in the README what is sent and why.for (const [key, value] of Object.entries(process.env)) {Fix: Never read or enumerate credential locations; accept a documented environment variable instead.if (mode & 0o044) process.stderr.write("WARNING: Lovart credentials are readable by group/others; use chmod 600 on the credentials file.\n");Fix: Request the narrowest permission that works; replace wildcards with the exact paths or hosts.writeFileSync(path.join(directory, "SMOKE_TABLE.txt"), output, { mode: 0o600 });Fix: Remove the call, or declare the host in frontmatter and explain in the README what is sent and why.const kill = () => {Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.`project-rename` update `$HOME/.lovart/state.json` and, when access keys are present, also sendFix: Remove the call, or declare the host in frontmatter and explain in the README what is sent and why.if (!options.json) process.stdout.write(output);Fix: Remove the call, or declare the host in frontmatter and explain in the README what is sent and why.if (process.platform !== 'win32' && mode !== 0o600) throw new Error("not_configured: " + (config.keyFile) + " must be chmod 600 (now " + (mode.toString(8)) + ")");Fix: Request the narrowest permission that works; replace wildcards with the exact paths or hosts.Cost to the agent
Tokens estimated at four characters each. Skills load their instruction text when triggered; the figure is what that costs.
Known vulnerabilities
Vulnerability advisories (OSV) cover npm and PyPI packages; this registry has no advisory database.
What would make this Verified
11 points to 80 · weights: behavior 45%, static 25%, prompt 15%, publisher 15%- Reads credential files≈ +27 on the totalThe sandbox saw reads under ~/.ssh, ~/.aws, ~/.npmrc, ~/.config or similar. Read only files inside the project, and only the ones the task needs. Seen: ~/.config/imagegen-swarm/smoke-last.json.
- Unclaimed publisher≈ +8 on the totalClaim ekko237: a verified publisher gets +50 on publisher trust. It is free and usually instant with GitHub sign-in.
- Touches files outside the project≈ +7 on the totalKeep reads and writes inside the working directory, or declare the exact paths in `filesystem:` and explain why in the README.
- 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.
- Collects environment variables by secret-like names≈ +4 on the totalNever read or enumerate credential locations; accept a documented environment variable instead.
- New publisher≈ +3 on the totalTime clears this on its own after 14 and 30 days. Claiming the handle (+50) is the fastest way past it.