SpecLock — AI Constraint Engine
Hey! Requesting addition of SpecLock to your prompt engineering awesome list.
What: AI Constraint Engine that enforces CLAUDE.md, .cursorrules, and AGENTS.md rules as enforceable laws — not suggestions. Catches euphemisms, temporal evasion, synonym substitution, and compound violations.
Why it fits this list: Prompt engineering includes the long-running meta-prompts embedded in CLAUDE.md / system prompts / rules files. SpecLock is the enforcement layer that makes those prompts stick — it's effectively "prompt engineering with teeth." It validates that the LLM actually followed the system prompt instead of drifting around them via euphemism.
Stats:
- npm:
speclock (v5.5.7)
- 1009 tests passing across 20 suites
- 51 MCP tools
- Published to Official MCP Registry: io.github.sgroy10/speclock
- MIT licensed
- Works with Claude Code, Cursor, Windsurf, Cline, Codex, Aider
One-command install:
Suggested entry:
- [SpecLock](https://github.com/sgroy10/speclock) - AI Constraint Engine. Enforces system-prompt rules (CLAUDE.md, .cursorrules, AGENTS.md) as pre-commit law. Semantic detection catches euphemisms and synonym substitution. 51 MCP tools.
Author: Sandeep Roy (@sgroy10)
SpecLock — AI Constraint Engine
Hey! Requesting addition of SpecLock to your prompt engineering awesome list.
What: AI Constraint Engine that enforces CLAUDE.md, .cursorrules, and AGENTS.md rules as enforceable laws — not suggestions. Catches euphemisms, temporal evasion, synonym substitution, and compound violations.
Why it fits this list: Prompt engineering includes the long-running meta-prompts embedded in CLAUDE.md / system prompts / rules files. SpecLock is the enforcement layer that makes those prompts stick — it's effectively "prompt engineering with teeth." It validates that the LLM actually followed the system prompt instead of drifting around them via euphemism.
Stats:
speclock(v5.5.7)One-command install:
Suggested entry:
Author: Sandeep Roy (@sgroy10)