Skip to main content
Agent skill

Frama C Proofreader

Verify and proofread C code with ACSL using this repository's Frama-C MCP workflow, Frama-C WP static proof, EVA alarms, and optional E-ACSL runtime checks. Use when asked to run Frama-C, check ACSL contracts, explain WP goals, inspect EVA alarms, validate C annotations, or state what a C proof does and does not establish.

Get started

Get the complete skill folder, including its scripts and reference files.

When to use this skill

Use Frama C Proofreader when an AI agent needs a reusable SKILL.md workflow for this job: Verify and proofread C code with ACSL using this repository's Frama-C MCP workflow, Frama-C WP static proof, EVA alarms, and optional E-ACSL runtime checks. Use when asked to run Frama-C, check ACSL contracts, explain WP goals, inspect EVA alarms, validate C annotations, or state what a C proof does and does not establish.

When not to use it

Skip Frama C Proofreader when the task is outside the coding category, or when a more specific skill in this directory already covers the same workflow with clearer triggers.

How to install

Where will you use it?

Install scope
Claude Code folder setup
mkdir -p ~/.claude/skills/frama-c-proofreader
# Save SKILL.md to ~/.claude/skills/frama-c-proofreader/SKILL.md
  1. Claude Code personal install: place the skill and its bundled files at ~/.claude/skills/frama-c-proofreader/SKILL.md. For a shared project, commit the folder at .claude/skills/frama-c-proofreader/ instead.
  2. Current Claude Code detects SKILL.md changes in watched directories during the session. If you just created a top-level skills directory, run /reload-skills. Check that the skill is listed before testing its trigger.
  3. For Claude chat or Cowork in the Desktop app, manage uploaded skills in Customize > Skills; copying files into a local Claude Code folder is not a Desktop install. Other agents may use different install locations.

Full install guide for Claude, Cursor, and Codex →

Try a starter prompt

After installing the skill, paste this starter prompt into your agent to verify it triggers on a small task.

Starter prompt
Use the Frama C Proofreader skill for this task: Use Frama C Proofreader when this documented workflow matches the task. Confirm any required inputs or files before making changes.

What supports this recommendation?

69/100 provisional content score. This measures documentation, practical detail and source signals. It is not a measured task success rate or a security certification.

Documentation
77/100
Practical detail
68/100
Source evidence
56/100
Source signals
72/100
Model compatibility evidence
  • claudeuntested

    No model-specific signal or recorded compatibility test was found.

  • chatgptuntested

    No model-specific signal or recorded compatibility test was found.

  • geminiuntested

    No model-specific signal or recorded compatibility test was found.

  • copilotuntested

    No model-specific signal or recorded compatibility test was found.

  • llamauntested

    No model-specific signal or recorded compatibility test was found.

  • perplexityuntested

    No model-specific signal or recorded compatibility test was found.

  • mistraluntested

    No model-specific signal or recorded compatibility test was found.

  • grokuntested

    No model-specific signal or recorded compatibility test was found.

Inferred compatibility is based on content signals. Check the app, its available tools and setup requirements before running a skill.

What this skill does

# Frama-C Proofreader

## Workflow

Prefer the MCP server when it is available. Start with the install check and the shortest useful call:

```text self_check {} reload_project {files, rte: true} check {function?, timeout?} ```

If `check` reports alarms or non-valid goals, inspect the concrete payload before changing code:

```text get_wp_goals {want: ["alarms"], function?, status?} get_wp_goals {want: ["investigation"], marker, depth} get_wp_goals {function, status?} context {function, want: ["current_annotations", "function_ast"]} ```

`investigation` takes the property marker off an alarm or goal row. A `check` payload is not proved while `incomplete[]` has any entry; read each entry's guidance before treating a valid goal as evidence.

Validate candidate ACSL with dry-run injection first:

```text inject_all_annotations {function, dry_run: true, annotations: [{kind, acsl, ...}]} inject_all_annotations {function, annotations: [{kind, acsl, ...}]} run_wp {functions: [function]} ```

Use `run_e_acsl` only as runtime evidence for concrete executions. If it is unavailable or fails to compile instrumentation, report that runtime checking was unavailable.

Read `references/mcp-workflow.md` for the full MCP call order, `references/wp-output.md` when interpreting WP goals, and `references/e-acsl-output.md` before making runtime-check claims.

## Reporting

Always separate these outcomes:

- Environment failure: tool missing, prover not configured, instrumentation failed. - Static proof gap: WP did not prove a goal; this is not automatically a bug. - Runtime violation: E-ACSL found a violation on an executed path. - Proven under assumptions: WP proved the generated goals for the ACSL and RTE configuration that actually ran.

Do not claim proof without tool output. Report exact goal names, proof counts, and whether `rte: true` or `-wp-rte` was used.

Before calling a function verified, run `check {files, smoke: true}` and confirm there is no `SMOKE_TEST_FAILED`: a contradictory `requires` or dead code makes every goal in scope provable, and only smoke tests notice.

## Boundaries

WP proves obligations generated from the C program and ACSL. It does not prove requirements that were omitted from the specification.

EVA alarms and WP goals are complementary. A clean WP result does not replace checking missing frame conditions or assumed callee contracts.

E-ACSL checks only executed paths and inputs. Passing one run is evidence, not exhaustive proof.

Intended uses

  • Use Frama C Proofreader when this documented workflow matches the task.

Implementation notes

  • Review the source instructions and adapt inputs before running the workflow.

Try it, then tell us how it went

A copied prompt does not tell us whether the task worked. Report your result after checking the output. We collect your tool and outcome, never your files or prompt.