agentleFS
tiago-oliveira-catiago-oliveira-ca/zenodotus-harness/skills/documentation/document-writer-reviewer/SKILL.md

document-writer-reviewer

Skill1 starsChanged 8 days agoMITView on GitHub

Review, rewrite, and normalize software delivery documents — BRDs, PRDs, FSDs/FRDs, NFRs, TRDs, ADRs, SADs, Epics/Stories, runbooks/postmortems, and developer handoff packages. Use when a document is incomplete, inconsistent, weakly structured, or needs to be upgraded into a cleaner, more traceable artifact. Trigger on: "review this document", "improve this doc", "rewrite this spec", "make this cleaner", "this needs polish", "review my BRD / PRD / FSD / NFR / TRD / ADR / SAD / epic / runbook", "audit this document", "what's missing from this spec", or any request to evaluate and strengthen a software delivery document.

We split this skill into the rules it gives an agent

10 rules. About 0 are common, 9 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈9

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use this skill to turn raw notes, uneven drafts, or weakly governed delivery documents into cleaner artifacts that teams can review, …

    line 15
  2. Distinctive:

    If the artifact type is unclear, infer it from scope and audience, then state that choice.

    line 48
  3. Distinctive:

    Call out missing information, unresolved assumptions, or governance gaps explicitly.

    line 56
  4. Very specific:

    Use shall for mandatory requirements and should for recommendations.

    line 61
  5. Distinctive:

    Keep business rationale separate from technical design and operational procedure.

    line 64
  6. Distinctive:

    Prefer tables, IDs, checklists, and clear headings over long narrative blocks when they improve execution clarity.

    line 65
  7. Distinctive:

    When surfacing review findings, classify each issue so readers know what to fix first:

    line 85
  8. Distinctive:

    When rewriting or reviewing:

    line 98
  9. Distinctive:

    keep edits implementation-ready rather than academic

    line 102
  10. Distinctive:

    use the most appropriate structure for the artifact instead of forcing one generic template

    line 103

≈ “Common” is an estimate: at least 1 in 5 of the 308,232 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Document Writer Reviewer

Use this skill to turn raw notes, uneven drafts, or weakly governed delivery documents into cleaner artifacts that teams can review, approve, and execute against.

Core Goal

Produce documentation that is:

  • structurally coherent
  • explicit about scope, ownership, and status
  • traceable across business, functional, technical, testing, architectural, and operational layers
  • precise enough to guide implementation or governance review

Artifact Router

Choose the review lens based on the document type:

  • BRD: Business value, scope boundaries, objectives, success measures, stakeholders
  • PRD: Product vision, personas, user stories, prioritized features, success metrics
  • FSD / FRD: User behaviors, flows, edge cases, functional requirements, acceptance criteria
  • NFR: Measurable performance, availability, security, scalability, compliance, and DR targets with verification methods
  • TRD / SRS: Technical constraints, interfaces, environments, verification methods, traceable requirement statements
  • ADR: Context, options, decision, status, and consequences
  • SAD / Blueprint: System structure, NFR coverage, integration, data, security, infrastructure, rollout
  • Epic / Story / Task: Traceability to a SAD component and requirement, acceptance criteria, Definition of Ready/Done, right-sized scope (INVEST)
  • Runbook / Postmortem: Executable procedures, alert-to-action mapping, escalation paths, blameless root cause, owned action items
  • Developer Handoff: Design assets, tokens, states, implementation notes, QA checks

If the artifact type is unclear, infer it from scope and audience, then state that choice.

Review Workflow

  1. Identify the artifact type, audience, and intended use.
  2. Preserve the original intent, but remove ambiguity, redundancy, and structural drift.
  3. Normalize the document into a clear section hierarchy with document control where needed.
  4. Tighten language so requirements and decisions are testable, traceable, and attributable.
  5. Call out missing information, unresolved assumptions, or governance gaps explicitly.
  6. Mark content you add beyond the supplied material as [Recommended Enhancement].

Writing Rules

  • Use shall for mandatory requirements and should for recommendations.
  • Replace vague language such as fast, secure, robust, or user-friendly with measurable criteria or named standards.
  • Keep business rationale separate from technical design and operational procedure.
  • Prefer tables, IDs, checklists, and clear headings over long narrative blocks when they improve execution clarity.
  • Preserve traceability between related artifacts when IDs or links already exist.
  • Execute every runnable command a document embeds before delivery, or mark it explicitly TODO-VERIFY: unexecuted — a copied "verify with:" command is a claim, not a fact, and its embedded commands are part of the document's testable surface. When the same command appears in multiple files, flag the duplication as a maintenance risk.

Shared Quality Checks

Review every document for these cross-cutting concerns:

  • Governance: version, status, ownership, approval path, auditability
  • Traceability: business need to requirement to design to test where applicable
  • Precision: one requirement or decision per statement where possible
  • Actionability: implementation, validation, or approval steps are clear
  • Completeness: assumptions, dependencies, risks, and exceptions are not hidden

Issue Severity Classification

When surfacing review findings, classify each issue so readers know what to fix first:

SeverityLabelDefinitionExample
Critical[CRITICAL]Blocks approval or creates compliance/legal riskNo approval path, undefined scope, regulation missed
Major[MAJOR]Significant gap that will cause rework or ambiguity in deliveryMissing acceptance criteria, orphan requirement, untraceable decision
Minor[MINOR]Structural or clarity issue that reduces quality but is fixable quicklyVague wording, inconsistent terminology, missing ID
Enhancement[Recommended Enhancement]Content added beyond original material that would improve the documentMissing risk section, suggested success metric

Surface Critical and Major issues before Minor ones. Never bury blockers in a list of nits.

Output Expectations

When rewriting or reviewing:

  • identify the document type
  • surface the highest-risk gaps first using severity labels
  • keep edits implementation-ready rather than academic
  • use the most appropriate structure for the artifact instead of forcing one generic template

Anti-pattern quick reference

Anti-patternInstead
Vague language (fast, secure, robust, user-friendly)Measurable criteria or named standards
Multiple requirements or decisions packed into one statementOne requirement or decision per statement
Business rationale mixed into technical design or operational procedureKeep the three layers separate
Long narrative blocks where execution clarity mattersTables, IDs, checklists, and clear headings
Blockers buried in a list of nitsSurface Critical and Major issues before Minor ones
Silently adding content beyond the supplied materialMark additions as [Recommended Enhancement]
Forcing one generic template onto every artifactUse the most appropriate structure for the artifact type
Rewriting away the original intentPreserve intent; remove only ambiguity, redundancy, and structural drift
Leaving the artifact type ambiguousInfer it from scope and audience, then state that choice
Hiding assumptions, dependencies, or risksCall out missing information and governance gaps explicitly
Delivering a doc with unexecuted embedded commandsRun each runnable command during authoring, or mark it TODO-VERIFY: unexecuted

Pre-flight checklist — run before delivering any document review

  • Scope, audience, and document purpose are explicit
  • Terminology is consistent and undefined acronyms are resolved
  • Requirements or decisions are atomic and testable where applicable
  • Ownership, status, and approval expectations are clear
  • Missing dependencies, assumptions, or risks are called out directly
  • The document is suitable for repository docs/ storage or formal review
  • Every embedded runnable command was executed during authoring or carries an explicit TODO-VERIFY marker

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

Reviews can't be read right now.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.