SLCR
An open standard for writing software requirements
FreeCLIOpen SourceAI Agents6 impressionslaunches Nov 31 comment
Comments
>log in to commentSpec-driven development is a leap forward, but over multiple sessions, agents drift from the spec and introduce bugs. SLCR lets you pin the software implementation against the spec, and give agents just the slice of the document they need, reducing context bloat. At Wack, we built SLCR to prevent multi-session drift and misalignment, and we're open sourcing the standard and the reference implementation.
SLCR is an open-standard format for sliceable, linked, checkable software requirements that agents can read and tools can validate.
- for
- Teams with long-lived codebases using AI agents for implementation.
- pricing
- open source
- license
- MIT
Key features
- Sliceable requirements — Break a spec into small DAG slices so each agent sees only its relevant requirement.
- Linked DAG structure — Define dependencies between requirements with AND, OR, XOR relationships.
- Checkable acceptance criteria — Attach any validator (tests, linters, ast-grep rules) to automatically verify requirements.
- CLI implementation — Rust-based slcr CLI to init projects, run checks, and integrate with CI pipelines.
- Agent skill integration — Provides a skill for agents to read SLCR and use the CLI during code generation.
- Human-readable Markdown output — Render the structured spec as a Markdown document for easy human consumption.
Use cases
- Validate that new API endpoints meet documented authentication requirements
- Ensure database schema changes match declared table specifications
- Run CI checks that confirm code passes all linked requirement validators
- Provide AI agents with only the necessary requirement slice for a task
SLCR vs alternatives
 SLCR |  Kiro |  Taskless |  LangChain |  Augment Code | |
|---|---|---|---|---|---|
| Best for | Spec-driven development with AI agents | Agentic IDE with spec-driven workflow | Enforcing agent instructions as rules | LLM app development framework | AI-assisted coding with deep context |
| Pricing | Open source | Subscription | Free | Free | Subscription |
| DevHunt upvotes | 1 | 0 | 58 | 0 | 0 |
| Launched | Nov 2026 | — | Oct 2026 | — | — |
- SLCR vs Kiro: Kiro focuses on an agentic IDE with spec-driven development, while SLCR provides a standalone requirement format and CLI.
- SLCR vs Taskless: Taskless turns markdown agent instructions into enforceable rules, whereas SLCR defines a structured requirement language.
- SLCR vs LangChain: LangChain is a framework for building LLM apps, not a requirement specification format like SLCR.
- SLCR vs Augment Code: Augment Code offers AI coding assistance with context, but does not provide a formal requirement spec or validation layer.
SLCR FAQ
How are requirements written in SLCR?+
Requirements are authored in JSON, TOML, or YAML files with an ID, title, body, modality, and acceptance criteria.
What tools can be used as validators?+
Any tool you declare, such as unit tests, linters, browser tests, or ast-grep rules, can be attached to acceptance criteria.
How does the SLCR CLI verify requirements?+
Run `slcr init` to set up, then `slcr check` which reports which validators pass or fail, suitable for CI integration.
Can SLCR be used with existing codebases?+
Yes, it is designed for long-lived codebases and works with existing test suites and validation tools.
Is SLCR an open standard?+
Yes, SLCR is an open standard with a reference implementation and a public GitHub repository.
Summarized by DevHunt from slcr.dev · Oct 29, 2026. Details may change; check the official site.
