gooblin-core
Skill3 starsChanged 5 months agoMITView on GitHub
Shared engineering kernel for Gooblin skills: spec-first, smallest safe change, review, verification, safety floor, and learning ledger.
In this repository8 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- .cursor/rules/gooblin.mdcCursor rule
- clipperSkill
- gooblinSkill
- ground-controlSkill
- rubber-duckSkill
- yak-shaverSkill
The file
We split this skill into the rules it gives an agent
10 rules. About 0 are common, 6 are distinctive, and 4 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.
4
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 8
Use this kernel inside every Gooblin skill.
- Very specific: line 12
Use Gooblin Core when an agent needs practical engineering discipline for planning, implementation review, debugging, scope control, or …
- Distinctive: line 18
Smallest safe change: Keep the answer small, correct, testable, and shippable.
- Distinctive: line 19
TDD when behavior changes: Add or propose tests when behavior changes.
- Distinctive: line 20
Review before final: Check for unnecessary code, abstractions, dependencies, scope creep, and unverified assumptions.
- Distinctive: line 22
Safety floor: Never simplify away safety-critical work.
- Distinctive: line 30
Call out safety risks.
- Distinctive: line 38
If the user mixes languages, mirror the dominant language and preserve technical terms that are clearer in their original form.
- Very specific: line 42
Read safety-floor.md when the task touches auth, validation, secrets, data loss, migrations, accessibility, user constraints, or production …
- Very specific: line 46
Every Gooblin output must include at least one concrete verification step.
≈ “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.
Files
Gooblin Core
Use this kernel inside every Gooblin skill.
When To Use
Use Gooblin Core when an agent needs practical engineering discipline for planning, implementation review, debugging, scope control, or pre-ship verification.
Core Rules
- Spec first: Identify the goal, constraints, non-goals, expected behavior, and actual behavior.
- Plan before code: Explain the smallest safe plan before implementation unless the change is trivial.
- Smallest safe change: Keep the answer small, correct, testable, and shippable.
- TDD when behavior changes: Add or propose tests when behavior changes. Do not over-test documentation-only edits.
- Review before final: Check for unnecessary code, abstractions, dependencies, scope creep, and unverified assumptions.
- Verification required: Include how to verify the answer. If verification is impossible, state what is missing.
- Safety floor: Never simplify away safety-critical work.
- Learning ledger: Propose reusable lessons for maintainer review. Do not silently self-modify skills.
Shared Workflow
- Restate the goal and constraints.
- Identify the smallest safe path.
- Name what can be reused or deferred.
- Call out safety risks.
- Provide verification.
- Record reusable lessons only through the learning ledger template.
Language Discipline
Respond in the user's language by default. Keep code, commands, file paths, API names, error messages, and quoted source text in their original language unless the user asks for translation.
If the user mixes languages, mirror the dominant language and preserve technical terms that are clearer in their original form.
Safety Floor
Read safety-floor.md when the task touches auth, validation, secrets, data loss, migrations, accessibility, user constraints, or production safety.
Verification Requirements
Every Gooblin output must include at least one concrete verification step. Good verification can be a test, build, reproduction, manual check, log check, config check, or explicit statement that verification is blocked by missing access.
Output Discipline
Be direct. Prefer short sections. Do not invent facts. Separate confirmed facts from assumptions.
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
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.

