| name | anti-slop |
| description | Detect and remove generic, unjustified, inflated, or unfinished work across code, UI, prose, research, and agent behavior. Use when creating or reviewing substantial output, when a result feels templated, or before declaring a user-facing task complete. |
Anti-Slop
Purpose
Raise specificity and usefulness by removing choices that are present because they are common defaults rather than because they serve the task.
Inputs
- The user's actual outcome and audience.
- Repository conventions, domain constraints, and existing content.
- The draft implementation or output.
- Verification evidence already collected.
Scope Boundary
Use this as a cross-cutting critique pass. It does not replace a domain review or implementation workflow; pair it with the skill that owns the code, UI, research, or writing task.
Process
- Restate the output's job in one concrete sentence.
- Identify elements that do not help that job: abstractions, decoration, claims, sections, dependencies, or process steps.
- Check for generic substitutions where project-specific evidence exists.
- Check for hidden incompleteness: placeholder content, untested paths, missing states, vague citations, or completion claims based on one command.
- Remove unnecessary elements before adding compensating explanation.
- Replace remaining generic choices with domain-specific ones supported by the brief or repository.
- Apply the relabeling test to UI and prose: if the structure survives unchanged when the subject is swapped, it is probably generic.
- Re-run the checks that prove the user-visible contract.
Constraints
- Do not equate minimalism with quality; necessary complexity should remain.
- Do not erase an established brand or project convention merely because it is common.
- Do not turn this skill into a universal checklist that blocks trivial work.
- Do not criticize subjective taste as an objective defect without evidence.
Verification
- Every prominent element has a task-specific reason.
- UI and prose contain subject-specific structure, not merely subject-specific nouns.
- No placeholders, invented claims, unexplained metrics, or decorative filler remain.
- The result covers relevant failure and edge states.
- Completion claims name the evidence that supports them.
- Removing an element would now reduce usefulness, clarity, or correctness.
Failure Modes
- Cosmetic cleanup that leaves an incorrect workflow untouched.
- Adding a new framework or design system to make output look intentional.
- Replacing one visual cliche with another.
- Polishing a generic template until it looks expensive without making it specific.
- Rewriting concise factual prose into a longer style guide performance.
- Declaring work clean because search found no
TODO markers.
Examples
UI: remove nested decorative cards, keep the operational grouping, and add the missing loading/error states.
Code: delete a pass-through abstraction and reuse the repository's existing helper.
Research: replace a universal ranking with separate measured dimensions and stated assumptions.
Sources
See ../../sources/anti-slop.sources.md.