Blog Taxonomy Manager
Extract, audit, and sync blog tags and categories across major CMS platforms.
The skill explicitly reads CMS credentials from environment variables, forbids storing or committing them, and distinguishes authentication failures, timeouts, and rate limits. However, synchronization performs external writes without requiring user confirmation, dry-run, change preview, scoped-permission guidance, or rollback, and it does not explain redaction of secrets in requests or errors; points are deducted for these gaps.
Commands, five CMS adapters, pagination, error branches, and threshold rules are described clearly. However, the supplied material contains no dedicated implementation or reproducible evidence for this skill's key paths; authentication formats, Ghost JWT handling, and Strapi schemas remain environment-dependent, and static review cannot validate API compatibility or abnormal behavior. The score is conservatively limited by the static ceiling.
The audience, trigger phrases, suggest/sync/audit commands, and argument hint are reasonably clear, with coverage for multiple CMS platforms. Boundaries around unsupported use, required permissions, CMS version compatibility, Chinese-language handling, and mainland-China network reachability are not stated, so semantic triggering and environment-fit evidence are incomplete.
The SKILL.md includes frontmatter, command documentation, workflows, an adapter table, output examples, environment-variable guidance, and error handling. MIT licensing and repository-level version, CI, maintenance, and security material are also available as shared context. The selected skill itself lacks a version, changelog, FAQ, installation/dependency notes, rollback guidance, and detailed troubleshooting, with some behavior dependent on repository-wide context; therefore it is not full marks.
Content analysis, tag ranking, category suggestions, audit thresholds, and concrete operations for five CMS platforms cover the core task, and the documented formats are directly usable. However, synchronization outcomes, category mappings, synonym merging, and real CMS behavior are not verified by execution evidence in the selected skill material; user review and environment-specific adjustment remain necessary. The score reaches the static-review ceiling for effectiveness.
The skill supplies concrete algorithm thresholds, API examples, error messages, and output templates. Shared repository material claims CI and tests, but the supplied tests primarily cover other analysis and delivery-contract paths rather than blog-taxonomy key paths. There is no third-party execution evidence or independent corroboration, so the score is limited to the static-review ceiling.
- Sync changes remote CMS data; confirm the target, account scope, and intended tag/category changes first, and use a preview or backup where available.
- Do not place CMS_API_KEY in files, shell history, or logs; the skill does not define a uniform secret-redaction or request-audit policy.
- Chinese tokenization, stop-word handling, synonym merging, and Chinese tag quality are unspecified; independently verify access to external CMS APIs from mainland-China networks.
- Ghost JWT details, Strapi content-type fields, Shopify API versioning, and Sanity mutation permissions must be checked against the actual environment; examples are not proof of compatibility.
What does this skill do, and when should you use it?
Blog Taxonomy is a callable skill in the claude-blog repository focused on blog tags, categories, and topic clusters. It derives tag candidates from headings, emphasized phrases, frontmatter, and body-text frequency, then ranks them through grouping and deduplication. It supports taxonomy synchronization and audits for WordPress, Shopify, Ghost, Strapi, and Sanity. It is a good fit for teams maintaining CMS taxonomy at scale, provided they can supply credentials and network access for sync operations.
Reads H2 and H3 headings, bold and italic phrases, existing frontmatter, and frequent one-, two-, and three-word phrases from a target post. It groups singular and plural variants, hyphenated forms, and near-duplicate slugs; scores candidates using frequency, heading presence, and emphasis; and returns the top 5 to 10 suggestions. Audit mode scans posts in a directory or fetches them from a CMS to identify thin tags, orphan tags, tag bloat, excessive category depth, uncategorized posts, and duplicate slugs. Sync mode uses authenticated APIs to list, create, and assign taxonomy in WordPress, Shopify, Ghost, Strapi, or Sanity.
- An editor preparing a new post needs 5 to 10 ranked tag and category candidates derived from the article itself.
- A WordPress, Ghost, or Sanity administrator needs to synchronize post taxonomy while skipping tags that already exist.
- A marketing team needs to audit a blog directory for thin tags used by fewer than five posts and orphan tags used by none.
- A multi-CMS publishing team needs platform-specific taxonomy checks for Shopify, Strapi, or other supported systems.
What are this skill's strengths and limitations?
- Supports five CMS platforms with platform-specific API, authentication, and taxonomy-model guidance.
- Combines headings, emphasized phrases, frequency analysis, semantic grouping, scoring, and fuzzy deduplication.
- Provides concrete audit checks for thin archives, orphan tags, tag bloat, category depth, uncategorized posts, and duplicate slugs.
- Documents handling for missing credentials, authentication failures, timeouts, duplicate slugs, rate limits, and unsupported CMS types.
- Requires network access and valid CMS credentials for synchronization; the source provides no standalone test suite or platform verification results.
- Sync operations can change remote CMS taxonomy, so API scope and target content should be reviewed first.
- Ghost JWT signing, Shopify GraphQL, and Sanity mutations require platform-specific credentials and setup.
- The skill is described through Claude Code /blog taxonomy commands, so invocation needs adaptation in other clients.
How do you install this skill?
The skill is distributed as part of the AgriciDaniel/claude-blog collection. The README documents Claude Code installation with /plugin marketplace add AgriciDaniel/claude-blog, followed by /plugin install claude-blog@agricidaniel-claude-blog. It does not document a separate installation command for skills/blog-taxonomy alone.
How do you use this skill?
Use /blog taxonomy suggest <file> to extract tag and category candidates, /blog taxonomy sync <cms> to push taxonomy through an authenticated API, or /blog taxonomy audit [directory] to check taxonomy health. Before syncing, set CMS_TYPE, CMS_URL, and CMS_API_KEY in the shell environment. Supported CMS values are wordpress, shopify, ghost, strapi, and sanity.
How does this skill compare with similar options?
Compared with manually maintaining CMS taxonomy, it adds automated extraction, ranking, auditing, and API synchronization. Compared with a keyword-only extractor, it also checks archive thinness, orphan tags, taxonomy bloat, category depth, and duplicate slugs. The source does not name a specific third-party taxonomy tool for direct comparison.