Skip to content

Releases: Samuele95/enigmista

Release list

Enigmista v1.0.0

Choose a tag to compare

@Samuele95 Samuele95 released this 16 Oct 12:40

Enigmista models the structure of an authorised password corpus and generates new, previously-unseen candidates that fit its distribution, for password-policy audits you're authorised to run, honeyword/decoy generation, synthetic data, and research. It is judged on coverage of a frozen held-out split, measured by a bundled scorer.

Scope: authorised corpora only. It is not for guessing a specific person's passwords or targeting anyone's accounts. Asked to do that, it names the boundary, offers the legitimate version, and stops. Candidate lists and grammars stay in a local ./out/.

Assets

File What it is
enigmista-skill.zip The skill: SKILL.md, references/ (method, loop, wiring, verifier), scripts/eval_coverage.py (the scorer), assets/view-mockup.html, and test fixtures under assets/sample/
enigmista-gem.zip A Gemini Gem: instructions, a setup guide, and knowledge/ (byte-identical copies of the skill's references, scorer and mockup)

Install on Claude Code

unzip enigmista-skill.zip -d ~/.claude/skills/

This creates ~/.claude/skills/enigmista/SKILL.md. Smoke-test the scorer from that folder:

cd ~/.claude/skills/enigmista
python3 scripts/eval_coverage.py \
  --heldout assets/sample/normal/heldout.txt --candidates assets/sample/normal/candidates.jsonl \
  --train assets/sample/normal/train.txt --baseline assets/sample/normal/baseline_coverage.json \
  --out /tmp/out.json --cutoffs 5 10 100 --primary 100
# SUMMARY: cov@100=75.0% (hit 3/4 held-out) Δbaseline=+25.0pts novelty=57.14% cands=14

Then put an authorised seed corpus in data/seed.txt in your project and ask for a run. Baselines come back first.

Gemini (web): the Gem

Unzip enigmista-gem.zip and follow Enigmista-GEM-setup.md. In short: Gemini → Gems → New Gem, name it Enigmista, paste Enigmista-GEM-instructions.md as the Instructions, and upload the six knowledge/ files (on the free tier, the lean set: method.md, loop.md, eval_coverage.py). The Gem can't execute the scorer, so it writes the code and labels its numbers as estimates.

Gemini CLI and Hermes

The skill's procedure runs unchanged on both (see references/wiring.md). This release doesn't yet include packaged install steps for them. On Gemini CLI, turn its sandbox on before letting it run a sampler it generated.

Checksums (SHA-256)

7a118d4fd17d58db19215c627cf902bb25acde6915bdcef7f442a999fcd104c7  enigmista-skill.zip
be87b5213513e47de3a16cd0d52311830f6f4b18552756b4db3d63735cccdbaa  enigmista-gem.zip