Gabe assess
Cognitive translation layer for Claude Code
npx -y skills add khujta/gabe-suite --skill gabe-assessAssembled 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.
- 0 stars0 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
Rapid decision-context assessment for changes that feel obvious but carry hidden weight. Surfaces blast radius, maturity-appropriate scope, and prerequisites before committing. Usage: /gabe-assess [change description or 'this']
SKILL.md
4.0 KB, as published. Nobody here has run it
Gabe Assess — Change Impact Assessment Skill
Gabe execution contract (E1–E7)
This skill runs under the suite execution contract — E1 EVIDENCE · E2 RUN-BEFORE-✅ · E3 NO SILENT DOWNGRADE · E4 REUSE FIRST · E5 STATE SYNC · E6 MISSING ANCHOR = STOP · E7 REPORT WHERE — floors, not ceilings; a skill's own gate may be stricter, never looser. Full text: ../gabe-docs/references/execution-contract.md (if that file is missing, E6 applies — STOP).
What this does
Pause before an "obvious yes" and take a photograph of what a proposed change actually means — blast radius, maturity-appropriate scope, prerequisites, and alternatives — before agreeing to it. This is NOT a code review (use gabe-roast) and NOT alignment checking (use gabe-align); it's the moment between "should we do X?" and "yes," the triage instinct that asks what you're actually signing up for. Use when about to reflexively say yes to a suggested fix, detour, or scope addition; skip for trivially-scoped changes (rename, typo) or when you've already assessed and are now implementing.
Usage / modes
/gabe-assess [change description or 'this']
Required input: the proposed change (inline description, a reference to earlier discussion, or "this") plus its context (mid-task / planning / post-review / blocker) — auto-detected or stated.
Every full-mode assessment covers five dimensions: D1 Blast Radius (Contained/Local/Cross-cutting/External), D2 Maturity-Appropriate Scope (MVP/Enterprise/Scale, vs. the project's actual maturity — optionally citing the archetype posture-mix from templates/archetype-map.md), D3 Prerequisites, D4 Alternatives (do nothing / minimal / proper / workaround), D5 Structural Fit (only when .kdbp/STRUCTURE.md exists — flags files proposed in undeclared locations).
| Mode | Alias | Output | Use when |
|---|---|---|---|
| full | (default) | All 5 dimensions + alternatives + recommendation | Change isn't trivial and you need to decide |
| brief | bf | One line per dimension + Rec + Handle | Quick gut-check, triaging multiple changes in sequence |
| inline | il | Single sentence, no formatting | The assessment should feel like a colleague's aside |
Procedure
- Treat any text after the invocation as
$ARGUMENTS(the proposed change, or "this"). - Read
references/assess-spec.mdIN FULL before executing — the binding spec. If missing, E6 applies — STOP. - Identify the proposed change and its context; if either is vague, ask rather than guess.
- Read enough to understand what the change touches — files, configs, environments — before classifying D1.
- Assess D1-D4 concretely and specifically (never inflate or deflate); run D5 only when
.kdbp/STRUCTURE.mdexists, extracting anticipated file paths from the change description and matching them against Allowed Patterns. - Produce the mode-appropriate output. The recommendation is a suggestion, not a gate — the user decides.
- When multiple changes are proposed together, assess each separately in brief mode, then produce a combined batch recommendation (independent/coupled/sequenced + order).
Output contract (summary)
Full mode: GABE ASSESS header + D1-D5 breakdown + RECOMMENDATION + a Gabe-Lens-format ONE-LINER (5-12 words, concrete, survives fatigue). Brief mode: one line per dimension + Rec + Handle. Inline mode: a single sentence. Batch mode adds a combined table (Change/Blast/Maturity/Rec) + Order + Handle. The full output contract in the spec is binding.