TheDocumentation Index
Fetch the complete documentation index at: https://mintlify.com/LilMGenius/polysona/llms.txt
Use this file to discover all available pages before exploring further.
$export / /export command reads the active persona and synthesises it into one or both of two portable configuration files: CLAUDE.generated.md for Claude Code projects and AGENTS.generated.md for Codex / OpenCode workflows. Once generated, these files can be dropped into any project root and immediately used by the target runtime — making your persona portable across all AI agent environments.
Invocation
$ARGUMENTS value is the target. Accepted values:
| Value | Effect |
|---|---|
claude | Writes personas/{active}/generated/CLAUDE.generated.md only |
agents | Writes personas/{active}/generated/AGENTS.generated.md only |
both | Writes both generated files |
| (empty) | Defaults to both |
What it does
Exporting to CLAUDE.generated.md (target=claude or target=both)
Builds personas/{active}/generated/CLAUDE.generated.md from persona.md and nuance.md:
- Work philosophy section — infers priority ordering, approach style, and risk stance from
persona.md. - Decision-making priorities — infers ordering and energy bias from
persona.mdcore data. - Tone rules — maps
nuance.mdvoice profile into concrete register rules and an avoid-list. - Context loading protocol — derives a startup context checklist from
persona.mdcore tags. - Anti-patterns to avoid — converts the
antifield inpersona.mdinto explicit do-not rules. - Adds a generated-file header note instructing the reader to review the file before copying it into a project root.
Exporting to AGENTS.generated.md (target=agents or target=both)
Builds personas/{active}/generated/AGENTS.generated.md as a Codex / OpenCode-compatible agent definition file:
- Includes agent name, description, and invocation method for each agent.
- Grounds the role split in the persona’s domain expertise from
persona.mdcore tags. - Keeps all entries portable across Codex-style and OpenCode-style
AGENTS.mdconventions. - Adds a generated-file header note indicating the file was auto-generated.
Prerequisites
The active persona must have completed interview and structuring steps:Output
Generated files are written to:generated/ directory is created automatically if it does not exist. The response includes the confirmed path(s) of every file written.
Guardrails & Notes
- Generated files contain a header note confirming they are auto-generated. Remove this header or adjust the content as needed before committing to a project.
$exportreads but never modifiespersona.mdornuance.md. Re-running it is always safe and produces a fresh synthesis from the current state of those files.- If the persona has changed since the last export (e.g. after additional
$interviewsessions), re-run$exportto regenerate files that reflect the updated persona. - The
generated/directory is inside the persona directory — it is scoped to the active persona. Switching the active persona and re-running$exportwill write to a differentgenerated/directory.
Related agents
$export has no dedicated agent — it synthesises from loaded persona files directly. The generated AGENTS.generated.md references the Profiler, Trendsetter, Content-writer, Virtual-follower, and Admin agents by default.