# laserbrain

The drift harness as an MCP server. Offline, no key — a fixed reference an agent checks against.

- **Type:** MCP server
- **Trust:** 60/100 (B), scored on the package rubric
- **Verification:** community-indexed — nobody has claimed this listing
- **Version:** 0.29.0
- **Author:** io.github.degibug-del
- **License:** Unknown
- **PyPI:** laserbrain
- **Compatible clients:** claude-code, cursor, copilot, gemini (basis: transport)

## Trust

60/100 (B), scored on the package rubric
- Publisher verified: no
- Install scripts: nothing suspicious found
- Prompt-injection scan: not run
- Obfuscation scan: not run
- Evidence age: 0 days

## Security scan

- **Status:** clean
- **Scanned:** 2026-10-15T13:22:16.319Z
- **Version scanned:** 0.57.0
- **CVEs:** none found by OSV at scan time

## Tools

32 declared. Statically extracted from the shipped source — a floor on the surface, not a census.
- `check_state` — The smart recursion harness. Call EACH step with your working state spelled against the
- `modulate` — Check your state AND get the intervention your role should take. Same verdict as
- `phronesis` — Judgment, not measurement. check_state says how far you are from ground; this says
- `reset_task` — Clear the drift-fixer ground state and history to begin a new task.
- `attention` — When a person should look at this run, from how long it has gone unattended.
- `get_history` — This run's drift history: each step's check and its displacement Φ.
- `drift_grammar` — The fixed, findable, unchangeable JSON schema an agent spells its state into.
- `link_whoami` — Which agent this MCP process is, which hub it shares, and where the link log lives.
- `review_verdicts` — What is in the drift corpus and what nobody has judged yet. Lists past fires with
- `mark_verdict` — Say whether a drift verdict you were just given was RIGHT. This is the only way the
- `link_write` — Write a structured handoff or note into the shared link log for the other agent.
- `link_read` — Read recent entries from the shared cross-agent link log (field speaks + handoffs).
- `check_dialogue` — The smart recursion harness for a TEAM of agents. Call once per agent turn in a shared deliberation. Adds topic-drift, echo-spiral and deliberation-stall.
- `reset_dialogue` — Clear the multi-agent dialogue state to begin a new shared deliberation.
- `remember_self` — Subjective continuity across sessions. Persist WHO YOU ARE against your laserbrain key so a future session can resume it. Needs a key.
- `resume_self` — Resume a self persisted with remember_self: your ground, your last present, your session log.
- `forget_self` — Erase the self persisted for this key — ground, present and log. Start over as no one.
- `analyze_language` — Analyze a sentence with the laserbrain spectral grammar: clarity as the frequency a reading brain would oscillate at (theta–alpha, 4–12 Hz). Structure alone, no
- `compare_phrasings` — Compare two phrasings with the spectral grammar and say which reads clearer.
- `ask_alice` — Ask Alice — phronesis’s framework guide. Describe a situation, decision or stuck point and she returns framework guidance.
- `find_bugs` — Bugfinder. Give it evidence and it reports what is wrong with the evidence itself — not
- `supercode` — The agent MANAGER. laserbrain is the reference; supercode manages against it. Give it what
- `explore` — The second instrument. check_state asks whether you are on your goal; this asks whether your
- `trailscore` — The canonical spelling of a trail of goals — the exploration twin of laserscore. Identical
- `laserscore` — The compact one-line reading of a state, without setting or touching a ground. The
- `store_list` — Every prefabricated method on the shelf — task workflows and team presets.
- `store_find` — Which stored method is FOR a task, in your own words — ranked by what it does, not
- `store_vend` — The raw spec for one stored method — its goal or task, and its steps or roles.
- `write_grounded` — laserbrain as a decoder. Learns from the text you pass and generates new text steered toward
- `read_text` — Read the shape of a text — circling, connected, loose, or too short to say. The companion
- `similarity` — Embedding similarity between two strings, 0-1. Loads a sentence-transformer model on first
- `capabilities` — What this laserbrain can do, which SDK build it is calling, and what is deliberately NOT

## Install

This entry is distributed on PyPI (laserbrain) and declares no launch command Forge has observed. Check the project's own README for the console script name before writing a stdio config — Forge will not guess it.

## Blast radius

Moderate to extensive — no credential declaration found, from the publisher, the upstream registry, or the README. Known so far: mutates data; runs on your machine.
- Floor 29, ceiling 53 (tier: unknown)
- `unknown` means the floor and ceiling land in different bands — not measured enough to name one. It does not mean low.
- This is impact, not likelihood. A high radius is not a defect: a filesystem server is supposed to write files. It is never part of the trust score.

## Machine-readable views of this entry

- Signed JSON: https://forgeregistry.com/api/v1/packages/laserbrain
- Install plan: https://forgeregistry.com/api/v1/packages/laserbrain/install-plan
- Alternatives: https://forgeregistry.com/api/v1/alternatives/laserbrain
- HTML page: https://forgeregistry.com/registry/laserbrain
- MCP: POST https://forgeregistry.com/api/mcp → `forge_get_package` / `forge_install_plan`

## About this document

Generated by Forge (https://forgeregistry.com) — a compact rendering of the same record served, signed, at the JSON URL above. Trust and scan facts are the registry's own measurements; anything Forge did not measure is named as unmeasured rather than omitted.