# Formal Spec Check

> >-

- **Type:** Skill
- **Install:** `agentstack add skill-daaa1k-skills-formal-spec-check`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [daaa1k](https://agentstack.voostack.com/s/daaa1k)
- **Installs:** 0
- **Category:** [Agent Skills](https://agentstack.voostack.com/c/agent-skills)
- **Latest version:** 0.1.0
- **License:** MIT
- **Upstream author:** [daaa1k](https://github.com/daaa1k)
- **Source:** https://github.com/daaa1k/skills/tree/main/skills/formal-spec-check

## Install

```sh
agentstack add skill-daaa1k-skills-formal-spec-check
```

Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent.

## About

# formal-spec-check

**USE FOR:** Alloy models beside docs | omission sweeps | bounded CLI + summaries | obligations tied to specs.

**DO NOT USE FOR:** Alloy tutoring only | code work without specs | teams refusing artefacts without rescoping.

Deep detail: [REFERENCE.md](REFERENCE.md).

## Preconditions

Locate **Alloy ≥ 6.2** runnable headlessly (`which alloy` **or** `java -jar org.alloytools.alloy.dist.jar help` listing `exec`/`solvers`). Stop if unresolved. Aim **≤ 60 s** per CLI invocation unless waived and explained in artefacts.

## Output norms

Default **Japanese** for chat prose + `basename.als.summary.md` readers; keep Alloy identifiers ASCII.

## Steps

1. **Sources:** cite repo Markdown/design paths first; pasted-only payloads land inside `formal/from-chat/YYYY-MM-DD-topic-slug/`.
2. **Traceability:** annotate `sig/fact/pred/check` constructs with concise `path + heading` breadcrumbs per REFERENCE.
3. **`check` posture:** obligations live in Alloy checks/asserts; optional `run` sketches only once justified in summaries.
4. **Temporal tier:** Alloy 6 trace/time features strictly after prose explicitly evolves over clocks; summarise horizon/`steps` limits.
5. **Ambiguity ladder:** catalogue `Assumptions (A…)` inside summaries—pause for stakeholder input whenever an assumption would flip verdicts.
6. **Runs:** whole-module `java -jar … exec` (or `alloy6 exec`) at `for 3 → 5 → 8` + Int widening in REFERENCE—stop after SAT, write required logs/summaries, note finiteness caveats on lingering UNSAT.

## Example prompts

- “Run Alloy on docs/features/auth/policy.md and summarise contradictions.”

CLI samples + triage: REFERENCE.md.

## Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [daaa1k](https://github.com/daaa1k)
- **Source:** [daaa1k/skills](https://github.com/daaa1k/skills)
- **License:** MIT

Install and usage instructions live in the source repository linked above.

## Pricing

- **Free** — Free

## Security capabilities

Automated source analysis of v0.1.0 — what this tool can access:

- **Network access:** no
- **Filesystem access:** no
- **Shell / process execution:** no
- **Environment & secrets:** no
- **Dynamic code execution:** no

*"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.*


## Versions

- **0.1.0** — security scan: passed — Imported from the upstream source.

## Links

- Listing page: https://agentstack.voostack.com/l/skill-daaa1k-skills-formal-spec-check
- Seller: https://agentstack.voostack.com/s/daaa1k
- Browse the marketplace: https://agentstack.voostack.com/browse

---
Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.