Naver Blog Publisher
v0.6.1Publish Naver Blog posts through the ACP marketplace flow (buyer-local thin-runner + sealed payload + offering execute). Use when a user asks to write/publis...
MIT-0
Security Scan
OpenClaw
Suspicious
medium confidencePurpose & Capability
The declared purpose (publish to Naver Blog from a local runner) aligns with the instructions, but the SKILL.md references several control-plane and auth variables (OPENCLAW_OFFERING_EXECUTE_URL, CONTROL_PLANE_URL + ACP_ADMIN_API_KEY, PROOF_TOKEN, SETUP_URL) that are not declared in the skill metadata. That mismatch between declared requirements (none) and runtime needs is concerning.
Instruction Scope
Runtime instructions tell the agent to run npx @y80163442/naver-thin-runner commands, start a local daemon, and auto-load X_LOCAL_TOKEN from ~/.config/naver-thin-runner/config.json. Those steps require reading a local config file and using tokens that are not listed as required env vars, and they suggest the agent will interact with local browser sessions and a local daemon — all sensitive actions beyond simple HTTP API calls.
Install Mechanism
There is no formal install spec; instead the SKILL.md instructs use of npx to run the scoped npm package @y80163442/naver-thin-runner. npx will fetch and execute code from the npm registry at runtime (user-scoped package), which is a moderate-to-high risk vector unless the package and publisher are verified.
Credentials
The skill metadata lists no required env vars, but the docs reference multiple environment/config items (OPENCLAW_OFFERING_ID, SETUP_URL, PROOF_TOKEN, OPENCLAW_OFFERING_EXECUTE_URL, CONTROL_PLANE_URL, ACP_ADMIN_API_KEY, X_LOCAL_TOKEN) and a local config path. Requesting or using unspecified secrets (ACP_ADMIN_API_KEY, X_LOCAL_TOKEN) is disproportionate and unclear.
Persistence & Privilege
The skill is not marked always:true and uses the platform's normal autonomous invocation setting. That alone is not a red flag, but combined with the above issues (local token access and remote package execution) it increases the potential blast radius if the fetched code or runtime steps are malicious.
What to consider before installing
This skill appears to be a local-publisher that expects a macOS runner and a one-time local Naver login, but the SKILL.md mentions several tokens/config entries that are not declared in the metadata. Before installing: 1) Ask the author for the source or homepage and for an explicit list of required env vars and config file formats. 2) Inspect the npm package @y80163442/naver-thin-runner source or release page — do not run npx blindly. 3) Verify whether the skill will actually require CONTROL_PLANE_URL/ACP_ADMIN_API_KEY (high-privilege) or will keep credentials strictly local. 4) If you must try it, run npx in a sandboxed environment and avoid providing global admin keys; keep user credentials on the local runner only. 5) Prefer a skill that declares its required env vars and provides an auditable install artifact (repository or release) rather than only npx instructions.Like a lobster shell, security has layers — review code before you run it.
latest
License
MIT-0
Free to use, modify, and redistribute. No attribution required.
SKILL.md
Naver Blog Publisher
Publish prepared content to Naver Blog from an authenticated local browser on the buyer machine.
Best for:
naver publishnaver blog publishkorean blog publish
First run:
doctor -> setup -> dry_run -> login once -> live
Production support:
macOS
Use when:
- a user already has final content and wants
naver publish,naver blog publish, orkorean blog publish - an OpenClaw agent needs a reliable publish endpoint, not a writing/SEO ideation tool
- the buyer can run a local Mac runner and complete one-time Naver login
Do not use when:
- the user needs the post drafted from scratch
- the environment cannot run a local runner
- the task is only research, SEO planning, or topic ideation
Inputs:
TITLEBODY- optional
TAGS - optional
PUBLISH_AT
Outputs:
- live publish:
naver_publish_result - preview:
dry_runresult with syntheticpublished_url - readiness check:
doctor/capabilitiesJSON
Recovery fields always expected on failure:
errornext_actionsetup_commandlogin_commandhintestimated_minutes
Runtime Config
OPENCLAW_OFFERING_IDdefaultnaver-blog-writerSETUP_URLorPROOF_TOKEN + SETUP_ISSUE_URLOPENCLAW_OFFERING_EXECUTE_URLpreferred- fallback:
CONTROL_PLANE_URL + ACP_ADMIN_API_KEY X_LOCAL_TOKENoptional and auto-loaded from~/.config/naver-thin-runner/config.jsonLOCAL_DAEMON_PORTdefault19090
Flow
doctor/capabilities- if
RUNNER_NOT_READY, run setup - run
publish_dry_run - if
login_required=true, run one-time login beforepublish_live publish_live
Commands
If the tool files are available, use them directly:
openclaw/skill-pack/naver-blog-writer/tools/doctor_capabilities
openclaw/skill-pack/naver-blog-writer/tools/publish_dry_run --title "Title" --body "Body" --tags "tag1,tag2"
openclaw/skill-pack/naver-blog-writer/tools/publish_live --title "Title" --body "Body" --tags "tag1,tag2"
One-time setup:
npx @y80163442/naver-thin-runner setup --setup-url "<SETUP_URL>" --auto-service both
Or proof-first:
npx @y80163442/naver-thin-runner setup \
--proof-token "<PROOF_TOKEN>" \
--setup-issue-url "<SETUP_ISSUE_URL>" \
--auto-service both
One-time login:
npx @y80163442/naver-thin-runner login
Start local daemon if needed:
npx @y80163442/naver-thin-runner daemon start --port 19090
Common Failures
RUNNER_NOT_READY
- setup has not been completed on this Mac yet
AUTH_EXPIRED
- complete one-time Naver login again before the next live publish
LOCAL_DAEMON_DOWN
- start the local daemon and retry
UNSUPPORTED_PLATFORM
- production support is currently macOS
Billing
publish_liveis the paid pathpublish_dry_runis a preview path and should not create a billable live publish
Notes
- This product is a publisher, not a writer
- keep
offering_id=naver-blog-writerfor compatibility - buyer credentials and session stay on the local runner machine
- advanced contract and schema docs live in
docs/ACP_CONTRACT.mdanddocs/OFFERING_SCHEMA.md
Files
1 totalSelect a file
Select a file to preview.
Comments
Loading comments…
