Home

percolation-inversion-compiler-ts Wiki

percolation-inversion-compiler-ts is an AI agent output checker and workflow report generator for Node.js.

It helps a JavaScript or TypeScript agent turn draft text, runtime files, local messages, and packet files into structured JSON reports. The reports show what was accepted, what can guide the next workflow step, what still needs checking, and what must not be treated as completed work.

In plain language, it answers:

• What did the agent claim?
• Was the input accepted by the checker?
• Can this report help the workflow continue?
• What evidence, identity, or verification work is still missing?
• What unresolved work must stay visible?
• Which next steps are only recommendations, not automatic execution?

This package is for npm and Node.js users. The Python implementation remains canonical.

• TypeScript repository: https://github.com/kadubon/percolation-inversion-compiler-ts
• npm package: https://www.npmjs.com/package/percolation-inversion-compiler-ts
• Canonical Python repository: https://github.com/kadubon/percolation-inversion-compiler
• Canonical Python wiki: https://github.com/kadubon/percolation-inversion-compiler/wiki

Fastest Start

Install and run a compact check:

npm install percolation-inversion-compiler-ts
npx pic-ts agent check --compact --text "Candidate packet: preserve residuals." --profile development

Plan the next workflow step:

npx pic-ts phase plan --compact --text "Candidate packet: preserve residuals." --profile development
npx pic-ts agent accelerate --compact --text "Candidate packet: preserve residuals." --profile development

Use pic-ts for npm and Node.js projects. The package also provides pic as a compatibility alias, but that name can be confused with the Python CLI.

What To Read First

• Quickstart: the shortest useful command path.
• Getting Started: install, run, and verify the package.
• Node Only Agent Loop: a Python-free local workflow for agents.
• How To Read The JSON: field meanings in practical terms.
• CLI Recipes: common command patterns.
• JavaScript SDK: root and subpath imports.
• Agent Messages and Packets: local message and packet sidecars.
• Safety and Limits: what the package does not authorize or prove.
• Compatibility With Python: how the npm package relates to the canonical Python project.
• Glossary: short definitions.

Practical Use Cases

• Check AI-generated text before reusing it.
• Keep missing evidence visible in CI or agent workflows.
• Create JSON reports that another agent can inspect.
• Export runtime reports as inert packet data.
• Inspect packet files without executing embedded command-like text.
• Build phase plans that show blockers and missing obligations.
• Validate schema and portability fixtures without Python.
• Keep accepted, workflow_usable, operationally_usable, and settled separate.

The Short Mental Model

agent output
-> structured JSON report
-> accepted or rejected envelope
-> residual ledger and missing obligations
-> recommended next workflow step
-> no hidden promotion to completed work

Useful output often has workflow_usable=true and settled=false. That is normal. It means the tool preserved unfinished work instead of hiding it.

Search Terms

Node.js AI agent output checker, npm AI workflow report generator, TypeScript JSON schema validation, LLM output validation, AI workflow verification, residual ledger, missing obligations, packet inspection, packet exchange, phase planning, runtime step report, JavaScript agent runtime, safe reuse checks, Python compatibility port, conformance fixtures, PIC, ECPT, BIT, TRC, SQOT, ALT.