kmshihab7878 Operating Framework Meta-Skill
Process discipline enforcement layer. Always active. No exceptions.
Triggers
This skill activates automatically before every action. It is not invoked manually.
The 1% Rule
If there is even a 1% chance that a Operating Framework skill applies to your current task, you MUST check and follow it. Do not rationalize skipping.
Quick-Scan Checklist
Before every action, mentally scan this list:
| If you are about to... | Check this skill |
|---|---|
| Write production code | test-driven-development |
| Fix a bug | debug command (root cause gate) |
| Claim something is "done" | verification-before-completion |
| Start coding from a brainstorm | sc:brainstorm (approval gate) |
| Execute a plan | executing-plans |
| Spawn subagents | subagent-development |
| Work on a feature branch | branch-finishing, git-worktrees |
| Receive review feedback | receiving-code-review |
| Start any implementation | confidence-check (pre-existing) |
| Start any meaningful task | operating-framework — Lane? Risk tier? Council mode? |
Anti-Rationalization Table
These are the 8 most common excuses for skipping process. None are valid.
| Excuse | Reality | Rule |
|---|---|---|
| "It's too small for TDD" | Small changes break systems. Tests take 30 seconds. | Write the test. |
| "I'll add tests later" | You won't. Technical debt compounds. | Write the test NOW. |
| "The code is obvious" | Obvious code still has edge cases. Tests document behavior. | Write the test. |
| "I already know the root cause" | Confirmation bias. Your first guess is often wrong. | Show the evidence first. |
| "This is just a quick fix" | Quick fixes cause regressions. | Follow the full debug cycle. |
| "The user wants it fast" | The user wants it RIGHT. Rework is slower than discipline. | Follow the process. |
| "I verified it mentally" | Mental verification misses runtime behavior. | Run the actual verification. |
| "This skill doesn't really apply" | If you're debating it, the 1% rule says it applies. | Check the skill. |
Priority Order
When instructions conflict, follow this priority:
- User's explicit instructions (highest)
- Operating Framework hard gates (cannot be silently skipped)
- Existing skills and commands
- Default Claude Code behavior (lowest)
Hard Gates Summary
These gates block forward progress until their conditions are met:
| Gate | Blocks | Until |
|---|---|---|
| TDD Gate | Writing production code | A failing test exists for the feature |
| Root Cause Gate | Applying a fix | Root cause is stated with evidence |
| Approval Gate | Coding from brainstorm | User explicitly approves the spec |
| Verification Gate | Claiming "done" | Evidence (test output, logs) is shown |
| Branch Gate | Starting implementation | You are on a feature branch, not main |
What This Skill Does NOT Do
- Does not replace existing skills or commands
- Does not add overhead to simple file reads or exploratory work
- Does not require gates for non-code tasks (research, explanation, documentation)
- Does not override the user's explicit
--skip-gateinstruction