Help us improve
Share bugs, ideas, or general feedback.
From skills-for-humanity
Writes and audits technical documentation, API docs, user guides, and specifications for completeness, sequence, precision, and audience calibration.
npx claudepluginhub human-avatar/skills-for-humanityHow this skill is triggered — by the user, by Claude, or both
Slash command
/skills-for-humanity:s4h-writing-technicalThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Technical writing fails when it is written for the person who already knows how the system works. The person who already knows doesn't need the document. The document is for the person who doesn't know — and that person's experience of the document is the only test that matters. The test is simple: can a reader with the assumed knowledge execute the task using only this document? If they cannot...
Drafts READMEs, API docs, tutorials, release notes, and reviews technical docs for clarity and structure. Activates on docs/ .md files and READMEs.
Author technical documentation through a repeatable workflow: audience analysis, Diátaxis-mode selection, structure, draft, review, publish. Covers READMEs, ADRs, API docs, and runbooks.
Enforces clarity, conciseness, and authenticity in technical writing for docs, guides, API refs; avoids AI patterns like 'delve into', 'leverage', 'robust'.
Share bugs, ideas, or general feedback.
Technical writing fails when it is written for the person who already knows how the system works. The person who already knows doesn't need the document. The document is for the person who doesn't know — and that person's experience of the document is the only test that matters. The test is simple: can a reader with the assumed knowledge execute the task using only this document? If they cannot, something is missing.
This is a higher bar than it sounds. Technical writers who know the system well have difficulty seeing the gaps — because for them, nothing is missing. The prerequisite knowledge is so obvious it doesn't seem worth stating. The step that requires understanding the system's mental model doesn't look like a step — it looks like common sense. These invisible assumptions are where most documentation fails.
The five failure modes in technical writing:
Audience miscalibration: Assumes knowledge the reader doesn't have, or over-explains what the reader already knows. Both are wrong. The first makes the document unusable; the second makes it condescending.
Incompleteness: Steps that require bridging knowledge the document doesn't provide. Often invisible to the writer because the knowledge is automatic.
Sequence errors: Prerequisites stated after they are needed, steps in the wrong order, troubleshooting information buried after the task it applies to.
Precision failures: Terms used inconsistently, ambiguous instructions ("configure the settings" — which settings?), passive voice that obscures who does what ("the file should be saved" — by whom, when, where?).
Missing examples: Non-obvious steps explained only in the abstract, without a concrete example that shows the reader what the correct output looks like.
Step 1: Audience Profile Who is the reader? What do they already know? What do they not know? What is their goal — and what is the fastest path from their current knowledge to task completion? Is the documentation calibrated to this profile, or to a different one (a more advanced user, or a more novice one)?
Framing check: Confirm the specific document and audience before continuing. State what you've identified — the actual document being audited, its subject matter, and the assumed reader profile — in one sentence, then use AskUserQuestion:
Step 2: Completeness Audit Read the documentation as if you are the assumed reader and have only their assumed knowledge. Walk through each task:
Flag every gap: name the missing knowledge, identify where in the sequence the reader would encounter it.
Step 3: Sequence Audit Check the order of information:
Step 4: Precision Audit For every instruction, ask: could a reader with assumed knowledge interpret this in two different ways?
AUTH_TOKEN")Step 5: Examples Audit Identify non-obvious steps — steps where the correct action or its output is not self-evident to the assumed reader. For each:
Before proceeding, use the AskUserQuestion tool. State your interpretation of the situation in 1–2 sentences — what is being analyzed and what the core question is — then ask:
Proceed based on their selection. If the user reframes, incorporate the correction before running any analysis.
Audience Calibration: [Who the assumed reader is / Whether the documentation matches that reader / Specific miscalibrations]
Completeness Gaps:
Sequence Issues:
Precision Flags:
Missing Examples:
Rewrite Suggestions: [Worst-offending sections rewritten with all issues addressed]
/s4h-writing-audience-calibration — audience profile is the anchor for every other decision; the calibration audit belongs here./s4h-writing-restructure — documentation structure often needs reordering, particularly moving prerequisites before the steps that need them./s4h-writing-line-editing — technical writing often suffers from passive voice and nominalisation; line editing is often the final pass.After delivering this output, use AskUserQuestion to offer the next move:
/s4h-writing-audience-calibration — Calibrate technical level for the audience/s4h-communication-clarity-audit — Audit technical clarity throughout/s4h-writing-restructure — Restructure technical content for clearer flow