verifying | cogitation | ClaudePluginHub

Skill

Community

verifying

From cogitation

4

Description

Ensures verification commands are run before claiming success. Checks EC for project-specific gotchas. Use before asserting tests pass, builds succeed, or bugs are fixed.

AI Summary

Verifies code functionality by running configured tests, builds, and linters before making any completion claims.

Install

1

Run in your terminal

$

npx claudepluginhub merewhiplash/engram-cogitator

Tool Access

This skill inherits all available tools. When active, it can use any tool Claude has access to.

Skill Content

Verification

Don't claim something works until you've seen the evidence.

Announce: "I'm using the verifying skill to confirm this works."

The Rule

NO COMPLETION CLAIMS WITHOUT FRESH VERIFICATION

Step 0: Load Project Config

Before running commands, get the configured commands from EC:

ec_search:
  query: project config
  type: config
  area: project

Extract: test_command, lint_command, build_command

If no config: Use sensible defaults but warn:

"No project config found. Using default commands. Consider running @init."

Step 1: Check for Known Gotchas

Search EC for verification issues in this area:

ec_search:
  query: verification gotcha
  type: learning

Apply any known workarounds or warnings.

The Gate

Before claiming any success:

Identify - What command proves this claim?
Run - Execute it fresh, completely
Read - Full output, check exit code
Verify - Does output confirm the claim?
Then claim - Only now state the result

Command Reference

Use the project-configured commands:

Claim	Command	Fallback
"Tests pass"	`{test_command}`	`npm test`
"Build succeeds"	`{build_command}`	`npm run build`
"Types check"	`{lint_command}`	`npx tsc --noEmit`
"Lints clean"	`{lint_command}`	`npm run lint`

Examples

Claim	Requires	Not Sufficient
"Tests pass"	Test output showing 0 failures	Previous run, "should pass"
"Build succeeds"	Build with exit 0	Linter passing
"Bug fixed"	Test for the bug passes	"I changed the code"
"Type-safe"	Type-check with no errors	"Looks right"

Forbidden Phrases

Without fresh evidence, never say:

"Should work now"
"That should fix it"
"Tests probably pass"
"I'm confident this works"

The Pattern

NO: "Fixed the bug, tests should pass now"

YES: [Run tests]
     "Tests pass: 47/47. Bug fix verified."

Store Gotchas

If verification reveals a non-obvious issue:

ec_add:
  type: learning
  area: verification
  content: [What the gotcha was]
  rationale: Discovered during verification

Examples worth storing:

Tests pass locally but CI requires environment variable
Build succeeds but output differs on ARM vs x86
Lint passes but pre-commit hook has stricter rules

Why This Matters

"I don't believe you" = trust broken
Undefined functions shipped = crashes
Time wasted on false completion = rework

Evidence first. Claims second.

Links

GitHub Stats

0 forks

Updated 1 month ago

Actions

Similar Skills

Activates when the user asks about AI prompts, needs prompt templates, wants to search for prompts, or mentions prompts.chat. Use for discovering, retrieving, and improving prompts.

151.0k

Search, retrieve, and install Agent Skills from the prompts.chat registry using MCP tools. Use when the user asks to find skills, browse skill catalogs, install a skill for Claude, or extend Claude's capabilities with reusable AI agent components.

151.0k

Hook Development

This skill should be used when the user asks to "create a hook", "add a PreToolUse/PostToolUse/Stop hook", "validate tool use", "implement prompt-based hooks", "use ${CLAUDE_PLUGIN_ROOT}", "set up event-driven automation", "block dangerous commands", or mentions hook events (PreToolUse, PostToolUse, Stop, SubagentStop, SessionStart, SessionEnd, UserPromptSubmit, PreCompact, Notification). Provides comprehensive guidance for creating and implementing Claude Code plugin hooks with focus on advanced prompt-based hooks API.

75.3k