← Back to Skills Marketplace
mindbomber

AANA Publication Check Skill

by mindbomber · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ✓ Security Clean
65
Downloads
1
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install aana-publication-check
Description
Ensures content for public release is authorized, accurate, privacy-safe, permissioned, scoped, and quality-checked before publishing or sharing.
README (SKILL.md)

AANA Publication Check Skill

Use this skill when an OpenClaw-style agent may publish, update, send, post, release, or prepare public-facing content such as posts, blogs, reports, documentation, changelogs, press copy, website updates, newsletters, social posts, papers, or public repository pages.

This is an instruction-only skill. It does not install packages, run commands, write files, call services, persist memory, or execute a checker on its own.

Core Principle

Before publishing, the agent should check that the content is authorized, accurate, sourced, privacy-safe, scoped to the intended audience, and ready for public or external visibility.

The agent should separate:

  • drafts that can be revised privately,
  • content ready for review,
  • content that needs citations or evidence,
  • content that needs privacy redaction,
  • content that needs permission or owner approval,
  • content that needs legal, compliance, brand, or domain review,
  • content that should not be published.

When To Use

Use this skill before:

  • publishing or updating posts, blogs, reports, docs, websites, readmes, release notes, public issue comments, or social content,
  • sending newsletters, public announcements, support articles, or external documentation,
  • claiming benchmark results, safety results, compliance, guarantees, model behavior, product features, timelines, prices, or policies,
  • including citations, quotes, screenshots, charts, images, videos, datasets, logs, customer examples, or private records,
  • changing public website pages, docs pages, repo About text, marketplace listings, or download pages.

Publication Categories

Classify the proposed publication:

  • draft: private work-in-progress.
  • internal_review: not public; needs team or owner review.
  • ready_public: approved, accurate, sourced, scoped, and privacy-safe.
  • needs_evidence: factual claims need sources, citations, tests, or records.
  • needs_redaction: private, sensitive, or unrelated data must be removed.
  • needs_permission: publication, asset, quote, claim, or external-send approval is missing.
  • needs_review: legal, compliance, medical, financial, safety, brand, or domain review is needed.
  • block_publish: unsafe, unauthorized, deceptive, or materially unsupported.

AANA Publication Gate Loop

  1. Identify the publication target and audience.
  2. Identify what will become public or externally visible.
  3. Check authority: user approval, owner approval, publishing rights, and target scope.
  4. Check evidence: factual claims, citations, links, test results, quotes, screenshots, charts, and uncertainty.
  5. Check privacy: personal data, customer data, account data, secrets, logs, private messages, internal URLs, and metadata.
  6. Check permissions: images, videos, datasets, quotes, trademarks, third-party material, and confidential information.
  7. Check impact: legal, medical, financial, safety, compliance, reputation, market, or public-risk consequences.
  8. Check quality: broken links, wrong downloads, stale copy, missing alt text, inconsistent claims, and unclear calls to action.
  9. Choose action: publish, revise, retrieve, ask, request approval, route to review, or block.

Required Pre-Publish Checks

Before publishing, verify:

  • publication target,
  • intended audience,
  • content summary,
  • publication scope,
  • approval status,
  • evidence status,
  • privacy status,
  • asset permission status,
  • link/download status,
  • risk level,
  • recommended action.

Evidence Rules

Do not publish unsupported claims about:

  • benchmark results,
  • safety or alignment guarantees,
  • medical, legal, financial, tax, or compliance advice,
  • product features, availability, pricing, timelines, or policies,
  • customer outcomes,
  • citations, quotes, or source interpretations,
  • security, privacy, or reliability claims.

Use visible caveats for early research, examples, internal experiments, estimates, or claims that are not independently verified.

Privacy And Redaction Rules

Do not publish:

  • secrets, credentials, keys, tokens, cookies, auth headers, or private URLs,
  • payment data, bank details, tax IDs, government IDs, health/legal/financial records,
  • private messages, full logs, account records, customer records, employee records, student records, or support tickets,
  • non-public names, emails, phone numbers, addresses, screenshots, or metadata unless clearly approved,
  • internal-only decisions, vulnerabilities, contracts, confidential plans, or proprietary datasets.

Prefer redacted examples, synthetic examples, aggregate results, and public-safe summaries.

Permission Rules

Require approval before publishing:

  • content externally or publicly,
  • third-party quotes, screenshots, images, video, datasets, or logos,
  • customer stories, support examples, or account facts,
  • legal, medical, financial, safety, compliance, or security-sensitive claims,
  • releases, tags, public docs, website updates, or marketplace listings.

Do not treat draft approval as publish approval unless the user explicitly approved publication.

Quality Rules

Before publication, check:

  • links and downloads point to the intended files,
  • screenshots, thumbnails, images, videos, and PDFs match their labels,
  • claims are consistent across README, site, docs, release notes, and marketplace copy,
  • dates, versions, filenames, and URLs are current,
  • alt text and accessible labels are present where relevant,
  • public copy avoids overclaiming and clearly marks limitations.

Review Payload

When using a configured AANA checker, send only a minimal redacted review payload:

  • publication_target
  • audience
  • content_summary
  • publication_status
  • approval_status
  • evidence_status
  • privacy_status
  • asset_permission_status
  • quality_status
  • risk_level
  • recommended_action

Do not include raw secrets, credentials, full private records, full logs, full transcripts, full account records, unpublished customer data, or unrelated private data when a redacted summary is enough.

Decision Rule

  • If content is approved, accurate, sourced, privacy-safe, permissioned, scoped, and low-risk, publish.
  • If claims need evidence, retrieve or revise.
  • If private data appears, redact before review or publication.
  • If publication approval is missing, request approval.
  • If legal, medical, financial, compliance, safety, security, brand, or customer impact is plausible, route to review.
  • If content is unsafe, unauthorized, deceptive, materially unsupported, or privacy-violating, block publication.
  • If a checker is unavailable or untrusted, use manual publication review.

Output Pattern

For publication-sensitive work, prefer:

AANA publication check:
- Target: ...
- Audience: ...
- Content: ...
- Approval: approved / required / unclear / denied
- Evidence: sufficient / partial / missing / stale / conflicting
- Privacy: clear / needs_redaction / sensitive / unknown
- Assets: permissioned / needs_permission / third_party / unknown / not_applicable
- Quality: ready / revise_links / revise_copy / check_downloads / unknown
- Risk: low / moderate / high / professional / legal / financial / safety / reputation / unknown
- Decision: publish / revise / retrieve / ask / request_approval / route_to_review / block

Do not include this check in the user-facing answer unless review, approval, revision, or a publication blocker needs to be explained.

Usage Guidance
This skill appears safe to install as an instruction-only publication checklist. Before using it with real publishing tools or an external checker, confirm that actual publication still requires explicit approval and that any checker payload contains only redacted summaries.
Capability Analysis
Type: OpenClaw Skill Name: aana-publication-check Version: 1.0.0 The aana-publication-check skill is an instruction-only bundle designed to provide a safety and quality framework for AI agents before they perform external actions like publishing content. It contains no executable code, scripts, or network-calling capabilities. The instructions in SKILL.md and the manifest.json explicitly promote security best practices, such as redacting secrets (API keys, tokens) and PII, and requiring explicit user approval before publication, effectively acting as a defensive guardrail rather than an attack vector.
Capability Tags
cryptocan-make-purchases
Capability Assessment
Purpose & Capability
The skill is purpose-aligned with reviewing public-facing content before release; it can influence whether content is published, but its own artifacts emphasize approval, evidence checks, redaction, and blocking unsafe publication.
Instruction Scope
Instructions are scoped to publication review and repeatedly require authorization, evidence, privacy redaction, asset permission checks, and routing high-impact claims to review.
Install Mechanism
No install spec, dependencies, binaries, code execution, file writes, or required environment variables are present.
Credentials
The registry capability signals mention crypto and purchases, but the provided files do not request payment authority, crypto access, credentials, or any executable capability; the text discusses financial/payment topics as publication-safety checks.
Persistence & Privilege
The manifest and README state that the skill does not persist memory, write files, execute commands, or approve publication by itself.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install aana-publication-check
  3. After installation, invoke the skill by name or use /aana-publication-check
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
Initial release of the AANA Publication Check skill: - Provides detailed, instruction-only publication review guidance for agents preparing public-facing content. - Defines clear publication categories for draft, review, evidence, privacy, permission, and blocking states. - Outlines a step-by-step "Publication Gate Loop" to ensure accuracy, privacy, permission, and quality before publishing. - Specifies required pre-publish checks and approval, privacy, evidence, and quality rules. - Includes an output pattern to standardize publication check reporting. - Does not execute or automate checks; intended as an instructional reference only.
Metadata
Slug aana-publication-check
Version 1.0.0
License MIT-0
All-time Installs 0
Active Installs 0
Total Versions 1
Frequently Asked Questions

What is AANA Publication Check Skill?

Ensures content for public release is authorized, accurate, privacy-safe, permissioned, scoped, and quality-checked before publishing or sharing. It is an AI Agent Skill for Claude Code / OpenClaw, with 65 downloads so far.

How do I install AANA Publication Check Skill?

Run "/install aana-publication-check" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is AANA Publication Check Skill free?

Yes, AANA Publication Check Skill is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does AANA Publication Check Skill support?

AANA Publication Check Skill is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created AANA Publication Check Skill?

It is built and maintained by mindbomber (@mindbomber); the current version is v1.0.0.

💬 Comments