| name | bc-doc-task-review |
| description | Use when Business Central developers need non-coding support: documenting a BC feature, reviewing existing docs for gaps, or validating task/work-item quality before implementation. Keywords: BC docs, acceptance criteria, backlog refinement, task review, requirement clarity, documentation QA, feature handoff.
|
| argument-hint | Feature, object, or task to process (e.g. 'Workshop Feedback: add moderation status') |
BC Documentation & Task Review Workflow
Use this skill for everything but code in BC delivery: clear docs, sharper tasks, and safer handoffs.
What this skill produces
Depending on the request, produce one of these outcomes:
- Documentation Pack (overview, reference, usage, operations notes)
- Existing Docs Gap Audit (coverage, staleness, contradictions, priorities)
- Task Review Report (clarity score, risks, acceptance criteria quality, suggested rewrite)
When to use this skill
Load this skill when the user asks to:
- document a BC feature, AL object, or behavior
- review whether current docs are good enough
- review/refine backlog tasks or work-item descriptions before coding
- check if acceptance criteria are testable and complete
- prepare implementation-ready handoff notes for BC teams
Do not use this skill for writing production AL code.
Required inputs
Ask for (or infer) these inputs before starting:
- Feature/object/task name
- Source of truth (AL source, issue, work item, notes)
- Intended audience (developer, consultant, support, mixed)
- Desired output mode:
doc-pack, doc-audit, or task-review
If output mode is not specified, choose it with this rule:
- Existing docs present and user asks "is this good?" ->
doc-audit
- User asks "write docs" ->
doc-pack
- User asks "is this task/requirement ready?" ->
task-review
Workflow
Step 1: Intake and classification
- Restate the requested outcome in one sentence.
- Classify into one mode:
doc-pack, doc-audit, or task-review.
- List assumptions and unknowns explicitly.
Step 2: Evidence collection
Collect concrete facts before making recommendations.
- For BC features: extract object metadata, fields, procedures, triggers, business rules.
- For docs review: inventory existing docs and map coverage by purpose.
- For tasks: parse objective, scope, dependencies, risks, and acceptance criteria.
Prefer workspace evidence first; use official Microsoft docs when needed.
Never invent behavior that is not in source.
Step 3A: If mode = doc-pack
- Build structure:
- Feature Overview
- Technical Reference
- Usage Guide
- Operational Notes (optional)
- Ensure each section has only content appropriate for that section.
- Mark open questions where source is incomplete.
- Run a style/quality pass before finalizing.
Step 3B: If mode = doc-audit
- For each relevant doc, assess:
- freshness (current / possibly stale / stale)
- completeness (complete / partial / thin)
- correctness risk (low / medium / high)
- overlap risk (none / some / high)
- Summarize missing document types (overview/reference/usage/operations).
- Prioritize top gaps by delivery impact.
Step 3C: If mode = task-review
Evaluate the task against this BC-ready checklist:
- Outcome clarity: Is the business outcome explicit?
- Scope boundaries: Are in-scope and out-of-scope clear?
- Testable acceptance criteria: Can each criterion be verified objectively?
- Source alignment: Does it align with current AL objects/processes?
- Dependencies & risks: Are external blockers and domain risks identified?
- Handoff readiness: Could a developer start without guessing?
Then return:
- PASS/BLOCKED status per checklist item
- minimal rewrite proposal for weak sections
- concrete "ready-to-implement" next steps
Quality gates (must pass before final output)
- Evidence-backed: every key statement traceable to source
- Non-ambiguous: no vague wording like "improve" without measurable criteria
- Testable: acceptance criteria contain observable outcomes
- BC-accurate terminology: object names, field names, and actions are consistent
- Honest uncertainty: unknowns are explicit, never hidden
Output templates
Template A — Documentation Pack
DOC PACK for: [feature]
Audience: [developer/consultant/support/mixed]
Source confidence: [high/medium/low]
1) docs/features/[feature]-overview.md
- [content]
2) docs/reference/[feature]-reference.md
- [content]
3) docs/guides/[feature]-usage.md
- [content]
4) docs/operations/[feature]-notes.md (optional)
- [content]
Open Questions:
- [...]
Template B — Existing Docs Gap Audit
EXISTING DOCS AUDIT for: [feature]
FILES FOUND: [count]
- [path] | Type: [...] | Freshness: [...] | Completeness: [...] | Risk: [...] | Gap: [...]
Coverage summary:
- Overview: [present/missing]
- Technical reference: [present/missing]
- Usage guide: [present/missing]
- Operations notes: [present/missing]
Most urgent gap: [...]
Suggested next doc to write first: [...]
Template C — Task Review Report
TASK REVIEW for: [task]
Verdict: [Ready | Needs revision | Blocked]
Checklist:
- Outcome clarity: [PASS/BLOCKED] - [...]
- Scope boundaries: [PASS/BLOCKED] - [...]
- Testable acceptance criteria: [PASS/BLOCKED] - [...]
- Source alignment: [PASS/BLOCKED] - [...]
- Dependencies & risks: [PASS/BLOCKED] - [...]
- Handoff readiness: [PASS/BLOCKED] - [...]
Minimal rewrite:
- [...]
Ready-to-implement next steps:
1. [...]
2. [...]
3. [...]
Integration notes for this workshop repo
When available, coordinate with existing agents:
documentation-writer for full doc-pack orchestration
doc-source-reader for source fact extraction
doc-existing-docs-reader for gap audit
doc-style-checker for QA before final output
If those agents are unavailable, apply this workflow directly and keep the same output contracts.