agentleFS
nesnilnehcnesnilnehc/ai-cortex/skills/review-technical-design/SKILL.md

review-technical-design

Skill8 starsChanged 2 months agoMITView on GitHub

Review an existing technical design against its modeling Spec, canonical quality Rule and applicable engineering Rule profiles before tasks or coding begin.

1 other project uses a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

6 rules. About 0 are common, 6 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈6

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    After technical design is drafted and before task derivation

    line 26
  2. Distinctive:

    After architecture, contract, data, operational or risk decisions change

    line 27
  3. Distinctive:

    Before coding a cross-boundary or quality-sensitive change

    line 28
  4. Distinctive:

    Do not invent project topology, budgets, SLOs or security classifications.

    line 46
  5. Distinctive:

    Do not restate engineering Rule text in the design.

    line 47
  6. Distinctive:

    Do not approve unresolved blocking decisions.

    line 48

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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

Skill: Review Technical Design

Purpose

Evaluate an existing technical design against technical-design-modeling, technical-design-quality and the engineering Rule IDs it claims to satisfy. This is an atomic pre-coding gate.

Core objective

Determine whether tasks can be derived without architectural ambiguity and whether applicable architecture, security, performance, observability, reliability and testing concerns have explicit tactics and verification.

Scope boundaries

This Skill reviews the design artifact and cited local context. It does not author the design, inspect implementation, run code review or aggregate downstream Skills.

Use cases

  • After technical design is drafted and before task derivation
  • After architecture, contract, data, operational or risk decisions change
  • Before coding a cross-boundary or quality-sensitive change

Behavior

  1. Load technical-design-modeling and technical-design-quality in full and record their versions, then load each engineering Rule set the design references.
  2. Resolve applicability per rule-modeling: whether the design triggers a quality attribute, a data change or an interface change, plus the project's profiles, parameters and valid waivers from .ai-cortex/config.yaml when present. For applicable UI changes, check decision traceability in the existing component, test and acceptance sections against the modeling Spec.
  3. Evaluate each applicable item independently, gathering only the evidence that item declares. An unresolved project value makes its item evidence-limited unless the design already records that absence as a blocking open question.
  4. For each triggered engineering concern, verify the design names Rule IDs, a system-specific tactic, its trade-off and a verification method; do not perform the future code review.
  5. Emit one finding per failed obligation, with category technical-design-quality, an exact section location, and the failed Rule ID in the description, for example technical-design-quality@<active-version>/TDES-007.
  6. Apply a waiver only when every waiver field is valid and its Rule ID and scope cover the exact finding. Report waived items separately.
  7. Return Rule coverage using the exact passed, waived, not_applicable and evidence_limited fields from the findings-list Spec.

Input and output

Input is one technical-design artifact plus its local parent and project configuration. Output follows findings-list: every finding uses category technical-design-quality and cites the failed Rule ID, and the coverage footer is metadata rather than a finding.

Restrictions

  • Do not invent project topology, budgets, SLOs or security classifications.
  • Do not restate engineering Rule text in the design; require references and design-specific decisions.
  • Do not approve unresolved blocking decisions.

Self-Check

  • The modeling Spec, quality Rule and referenced engineering Rules were loaded.
  • Applicable profiles and parameters were resolved.
  • Conditional quality-attribute design and applicable UI decision traceability were checked.
  • Findings are precise, category-compliant and each cites one failed Rule ID.
  • Every applicable Rule ID received a pass, a finding, a valid waiver or an explicit evidence limitation.
  • The coverage footer distinguishes passed, waived, not applicable and evidence limited.
  • No implementation review was performed.

Examples

Example 1: new public API

The design defines endpoints but no compatibility tactic or contract verification. Emit a finding against the quality-attribute design and cite the missing public-contract Rule references.

Example 2: no database change

Accept an explicit “no database change” statement. Do not require a migration plan when its trigger is false.

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.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

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.