Write and revise developer blog posts or run a humanizer / humanization pass on existing drafts using the AI anti-pattern catalog, structural audits, and reusable personal and explicitly selected corporate writing identities.
72
91%
Does it follow best practices?
Run evals on this skill
Adds up to 20 points to the overall score
View guide
Low
Low-risk findings worth noting
This skill is an action router — pick the step that matches the user's intent and execute only that step. Do not run other steps; do not parallelize.
Corporate identity owns audience, brand values, terminology, evidence standards, approved
claims, positioning, and editorial constraints. It does not prescribe an individual
author's personality. Before any action, read
skills/blog-writer/references/identity-spec.md completely and follow its package,
provenance, strength, ownership, approval, and storage contracts.
For Step 1, also read the shared storage contract before writing:
skills/blog-writer/references/identity-storage.mdEvery package has an identity.json manifest, concise shared guidance and routing in
identity.md, and source authority, dates, contradictions, and consequential inferences in
sources.md; optional Markdown resources are declared by role in the manifest.
Use one approval loop for every consequential create or update: set status to draft,
show the one-paragraph summary, Required and Avoid guidance, high-impact inferences, and
unresolved items, then ask one focused question at a time. If the user rejects or corrects
anything, revise the package, trace each changed conclusion to its cited sources, and present
it again. Set status to approved only after explicit approval.
Inventory every source the user supplies before drawing conclusions. Accept local files or directories, URLs, pasted text, skills, style guides, example content, product documentation, messaging frameworks, and editorial feedback. Establish each source's scope, authority, and relevant date or version. A current formal guide normally outweighs an older example, but conflicting authoritative policies remain unresolved until the user decides.
Compile evidence about intended audiences, reader relationship, values, trust signals, tone boundaries, terminology, product positioning, sources of truth, evidence standards, format-specific conventions, editorial checks, and stated legal or reputational constraints. Keep explicit policy separate from patterns observed in examples. Compile the identity and review guidance from source skills without copying their workflow.
Resolve the destination through the shared storage contract. Put concise shared guidance
and routing in identity.md; add roles such as brand, terminology, editorial-review,
and product-context only when the evidence supports them. Record every source and
consequential inference in sources.md. Run the approval loop, then finish here.
Read the existing package and its provenance before reading new evidence. Preserve valid guidance and source history. Identify which conclusions the new sources confirm, weaken, replace, or leave unresolved; never regenerate the package as if prior decisions did not exist.
Apply accepted changes through the same package contract as Step 1. Keep the existing status only when changes do not alter consequential guidance. For consequential changes, run the approval loop. Finish here.
Compare the package and its provenance against the current authoritative sources and recent representative content the user names. Report stale, unsupported, contradictory, overfit, and missing guidance with the evidence for each finding. Pay special attention to old examples outweighing current policy and unsupported claims presented as approved language.
Do not modify the identity during an audit. If the user later asks to apply accepted findings, that follow-up request routes to Step 2. Finish here.
.tessl-plugin
skills
blog-writer
cfp-abstract
create-corporate-identity
create-personal-identity
references