Probaboracle

probably a mini oracle. definitely a mini chatbot

Probaboracle is a small, local, agent-backed CLI mini chatbot using the Polinko research model.

It only accepts four question types:

• what
• when
• why
• where

That narrow surface is the point. Probaboracle is not trying to be a general chat tool. It is a small instrument for studying whether a model can stay coherent, vague, answer-shaped, and product-specific inside tight interaction guardrails.

Current research stage:

• clean baseline reset
• source state: proper-config candidate
• next eval gate: eval-pulse

Most recent diagnostic snapshot:

• Research Beta 6.0
• method: fail-pressure pulse
• first fixed-prompt pulse: FAIL
• result: 1 anchor / 13 counted seams / 0 excluded
• status: diagnostic snapshot, not a clean baseline

Most recently closed beta:

• Research Beta 5.1
• retain + evict

In this repo, a new beta gets pinned when the method change alters what the evidence means, not just when wording or procedure gets tidier. The current reset keeps the eval-pulse method but stops treating the first Beta 6.0 line as live baseline proof.

The next baseline starts from structural config only: fixed prompt types, binary verdicts, runtime settings, and a minimal routing prompt in the agent path. Earlier prompt scaffolds and config-level word lists are kept out of the fresh line. The next comparable evidence should be a new fixed-prompt pulse, run one prompt at a time after the rate-limit / prepaid-credit boundary is healthy.

What This Repo Demonstrates

• constrained one-node generation through a fixed prompt surface
• structural config without prompt phrase banks
• coherence-first evaluation instead of one overloaded product verdict
• beta-specific evidence gates:
  • row-level baselines use PASS / FAIL
  • closed Beta 5.1 failures can then earn RETAIN / EVICT
  • the Beta 6.0 snapshot judged one fixed-prompt pulse at a time
  • pulse rows are only evidence inside that pulse:
    • anchor
    • counted_seam
    • excluded_noise
  • the pulse receives one PASS / FAIL verdict
  • the clean-baseline candidate keeps the same pulse method before the next beta-boundary decision

Run It

make install
probaboracle

The app opens a compact terminal loop. Choose a question type with the arrow keys, press enter, or hit esc to exit.

The operator commands, eval workflow, and setup checks live in the runtime runbook.

Core operator commands:

make start
make end-preflight
make end-git-check
make caffeinate-status
make decaffeinate
make check

Read Next

• docs/research/README.md
  • beta map and research reading path
• docs/research/070_CB-CLEAN_BASELINE_RESET.md
  • current reset boundary, docs cleanup, and first local pulse plan
• docs/runtime/templates/README.md
  • public templates for future research docs and pulse reports
• docs/governance/DECISIONS.md
  • durable runtime and eval decisions

Licence

Apache-2.0. See licence.

---

Probaboracle is not a resource for yes, no, or maybe.