Setup pstack

Write each pstack skill's references/models.md, the per-skill override file that sets pstack's model per role. Every skill that reads role→model config has one. The skills read their own file and fall back to the parent model when a line is absent, so this is an override layer, not a requirement.

Steps

1. Locate the per-skill config files

Find pstack's skills wherever they are installed: this repo's skills/, ~/.prime/agent/skills/, or node_modules/<pkg>/skills/. Every skill with a references/models.md is configurable. Read each file; its header comment documents the role lines and their legal values for that skill.

2. Detect available models

Resolve the model selectors you can pass to a child in this session with await rlm.find_models(...); that is the dependable source. If you cannot detect any, ask the user to paste the selectors they have access to. Never write a real selector you have not confirmed is available. The alias inherit-parent is always valid even though it is not a detected selector.

3. Load current state

For each skill, the current choices are the role lines present in its references/models.md. A file with no lines (or no file) means every role falls back to the parent model.

4. Map and confirm

Show every role with its current model, marking any real selector not in the detected set as needing a choice. Ask the user whether to accept as-is or change specific roles, offering the detected models plus inherit-parent (this role runs on the parent chat model) as the options. For panel roles (arena runners, how critics, architect runners, interrogate reviewers) the value is a list, and one subagent runs per entry, inherit-parent entries included, so the list length sets the count. arena cross-judge pool is also a list, but Arena selects one value from it whose model family differs from the parent's when possible. swarm workers is the default model for every worker unless a race or comparison assigns another model per arm. Follow each skill's own file for its exact role names and list semantics.

5. Validate

Every real selector written must be in the detected set; inherit-parent always passes. If a chosen real selector is not available, stop and ask again. A file pointing at a model the user cannot use breaks every delegation that reads it.

6. Write each file

Write each references/models.md with one line per role, using the same labels the skill's own file documents. Keep the file's header and example comments as-is; overwrite the values so re-runs stay idempotent. The reflect file shows the canonical shape:

# reflect model configuration

One line per role. Delete a line (or this file) to inherit the parent model.

# judgment: <model selector>
# tooling: <model selector>

Resolve a valid selector with await rlm.find_models(...) and pass the exact returned selector as model= when spawning. inherit-parent or omitting model runs the child on the parent chat model.

7. Confirm

Tell the user the files were written and that the choices apply to new runs. Re-running this skill updates them.

8. Offer a verification skill (optional)

Check whether the project has a way to drive the real app for proof (a verify-* skill, or an existing harness). If not, offer once: "want a project-local verification skill, so agents can drive the app the way a user does and prove changes work? I can generate one with /create-verification-skill." On yes, invoke /create-verification-skill. On no, move on without pushing.