fla-kda

FLA KDA kernel workflow and public technical notes. Use when modifying or reviewing fla/ops/kda/**, KDA gate modes, chunk intra/inter kernels, safe_gate behavior, KDA backends, or KDA-specific tests and benchmarks.

Quality

81%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Quality

Content

75%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A well-structured, technically dense reference that assumes competence and provides concrete, codebase-specific guidance with a validation checklist. The main room for improvement is trimming the numerical derivation and adding an explicit validate-fix-retry loop.

Suggestions

Tighten the "Safe gate numerical note" section: the midpoint-offset derivation could be condensed to the invariant (local offset per exponentiation) without the full base-2 unit walkthrough.

Add an explicit validate-fix-retry loop after the correctness checklist, e.g. "run fla-correctness-coverage -> if any axis fails, fix and re-run until green".

Consider extracting the lengthy "Public code map" API listing into a references file so the SKILL.md overview stays leaner, signaling it one level deep.

Dimension	Reasoning	Score
Conciseness	Dense and technical, assuming Claude's competence (no explaining of FLA/KDA/exp2 basics); every section is targeted, though the "Safe gate numerical note" derivation runs somewhat long.	4 / 5
Actionability	Concrete API names, file paths, exact gate formulas, and precise constraints ("-5 <= lower_bound < 0", specific fla.utils helpers) give mostly executable guidance; no runnable code snippets, but that is justified for an instruction-only skill about an existing codebase.	4 / 5
Workflow Clarity	The "Correctness checklist" plus the explicit "Do not change one path without checking the other" warning provide a clear validation sequence with most checkpoints present, though there is no explicit validate-fix-retry feedback loop.	4 / 5
Progressive Disclosure	Well-organized into clear section headers (Public code map, Gate modes, Safe gate numerical note, etc.) with no bundle files; content is appropriately sectioned in a single file, with only minor organization gaps.	4 / 5
	Total	16 / 20 Passed

Description

87%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A strong, specific description that explicitly states both what the skill covers and when to use it, scoped tightly to the FLA KDA kernel code path. The only mild weakness is that the action verbs are limited to "modifying or reviewing".

Dimension	Reasoning	Score
Specificity	Names the domain and lists several specific target areas ("KDA gate modes, chunk intra/inter kernels, safe_gate behavior, KDA backends, or KDA-specific tests and benchmarks"), though the verbs are limited to "modifying or reviewing".	4 / 5
Completeness	Clearly answers both "what" ("FLA KDA kernel workflow and public technical notes") and "when" ("Use when modifying or reviewing fla/ops/kda/**...") with concrete trigger phrases.	5 / 5
Trigger Term Quality	Good keyword coverage for this niche ("fla/ops/kda/**", "safe_gate behavior", "KDA backends", "tests and benchmarks") that a developer in this codebase would naturally say, with a few natural variations missing.	4 / 5
Distinctiveness Conflict Risk	Clear niche tied to a specific code path (fla/ops/kda/**) with distinct triggers; minimal conflict risk with other skills.	5 / 5
	Total	18 / 20 Passed

Validation

100%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: fla-org/flash-linear-attention
Path: .agents/skills/fla-kda/SKILL.md
Commit: 27967b9

Reviewed: 3 days ago

Table of Contents

Discovery Implementation Validation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.