Constraint review
Skill iplaylf2/coding-profile/profile/agent/skills/constraint-review
Personal coding profile kit for reusable defaults and setup
npx -y skills add iplaylf2/coding-profile --skill constraint-reviewAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
2 things to look at
- no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
- 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Use when the user needs to evaluate, decide, or change how a supporting constraint or defensive measure in code or prose serves an artifact's primary responsibility, including its validity, scope, disposition, or owning boundary. Do not use when the relevant constraint decisions—including validity, scope, behavior, disposition, required outcome or reader effect, and owning boundary—are settled and the request only implements or edits the constraint, or when the request only makes a structural placement decision after responsibilities are settled.
SKILL.md
7.6 KB, as published. Nobody here has run it
Constraint Review
Review Boundary
This skill determines whether and how an in-scope constraint should serve an artifact's primary responsibility and which boundary should own the responsibility for the required outcome.
Treat decisions settled by authoritative requirements or contracts as inputs; evaluate only the validity, scope, behavior, disposition, and ownership left open.
For prose, treat qualifications, caveats, exceptions, and warnings as constraints on what the reader may infer or do.
Context And Responsibility Questions
Use requirements, schemas, public contracts, threat models, observed behavior, tests, and local conventions as evidence. Distinguish a governing contract from a current observation or copied assumption.
- What primary result does the artifact owe, to whom, and through what primary path?
- Which exact constraint and named scope are under review?
- Which request inputs are binding requirements, which are goals, preferences, rationales, or candidates, and which validity, scope, behavior, disposition, or ownership decisions remain open?
- Which evidence governs the open decisions, how authoritative and stable is it, and how confidently does it support the claimed responsibility?
- What material invariant, trust boundary, failure-containment duty, resource need, or audience need establishes a responsibility, which boundary owns it, and what outcome must that boundary provide?
- What trigger makes that responsibility relevant, and what structural, contractual, observed, or adversarial evidence establishes its reachability and material consequence?
- What observable outcome, reader conclusion, valid behavior, and cost follow with and without the constraint?
- Which responsibility does another mechanism already fulfill, and what outcome does it provide? What independent obligation remains at this boundary, and what local outcome does it require?
- Does the constraint fit the responsibility in trigger, behavior, scope, wording, duration, and cost? If the responsibility is transitional, what observable exit or reevaluation condition applies?
- What unresolved input or missing authority could materially change an open decision or risk, and can it be inferred or obtained before asking a focused question?
Code Questions
- Which input, state, timing window, dependency failure, or trust transition triggers the responsibility, and what success, failure, recovery, or uncertainty outcome does the boundary owe?
- Which response strategy—for example, rejection, recovery, degradation, retry, delegation, or surfacing failure or uncertainty—produces that outcome, and which resulting states must callers distinguish?
- When behavior is controlled or enforced at multiple boundaries, what independent obligation does each boundary own, and what local outcome does that obligation require from the enforcement?
- When a local rule mirrors behavior owned elsewhere, should it follow that owner or preserve a distinct local contract? What governs that relationship, and how would drift become observable?
- Would a fallback, retry, or degradation preserve valid work, keep success within its governing contract, and preserve each required failure, recovery, or uncertainty distinction?
- As assumptions, resources, or timing change, would the rule fail open or closed, misclassify a state, or change the valid state or time window, and what capacity or lifecycle does that behavior protect?
- Can the existing model, type, contract, or source of truth express the responsibility and its distinct outcomes without duplicated state or logic?
Prose Questions
- Which claim does the qualification limit, and what reader inference, decision, or action changes with and without it?
- Where must the qualification appear so it shapes the reader's understanding before a decision or action relies on the claim?
- Which supplied wording or terminology is an exact requirement, and which instead indicates the intended boundary or consequence to express in local prose?
- Can the primary rule express the responsibility more directly while keeping the owning boundary and concrete consequence clear?
Disposition Questions
Use the relationship among evidence, responsibility, owner, required outcome, and constraint expression to consider:
- Keep when the current constraint, responsibility, and owner fit.
- Clarify when the responsibility and owner fit but the expression is imprecise.
- Narrow when the constraint exceeds the responsibility.
- Move when the responsibility is valid but another boundary owns the responsibility for the required outcome.
- Remove when no independent obligation at the current boundary requires the constraint's local outcome, or when the constraint is disproportionate to that obligation and outcome.
- Add when governing evidence establishes an unprotected, in-scope responsibility at a boundary that can produce the required outcome.
Then decide:
- Which disposition or combination most directly and proportionately satisfies the evidenced responsibility?
- Should the responsibility be carried by behavior, a prose qualification, or both?
- Which candidates satisfy the governing inputs and responsibilities? If none do, which conflict must be resolved before implementation?
- Among valid candidates, how should stated goals and preferences determine the choice?
Implementation And Validation
When implementation is requested, make the coherent code, prose, and test changes implied by the settled disposition. Within the owning boundary, prefer one explicit source of truth; retain enforcement at multiple boundaries only when each boundary owns an independent obligation that requires a local outcome from that enforcement.
Run applicable existing checks for the primary path, constraint trigger, and required outcome or reader effect. Test observable contracts rather than the mere presence of a branch or phrase. Add or adjust a focused check only when the implementation changes an otherwise uncovered contract.
For prose, reread the primary path with and without the constraint, check required terminology and nearby claims for contradiction, and run existing documentation checks when they cover the affected text.
Confirm that each retained or introduced constraint maps to an evidenced responsibility, and each added validation maps to a changed or otherwise uncovered contract.
Result
Lead with the disposition and decisive evidence. If no valid disposition is available, lead instead with the governing conflict and the decision needed before implementation. State how the constraint fits the evidenced responsibility, the material outcome or reader conclusion with and without it, and which current or target boundary owns the responsibility and must produce the required outcome. Identify the governing requirement or responsibility that controlled any material tradeoff.
Report any implementation, the exact validation performed, and its result. Distinguish executed checks from conceptual review, and report residual uncertainty only when it still affects the decision or risk.