Generalize seats beyond harness identities #123

New issue

Closed

opened 2026-07-29 04:39:41 +00:00 by coilyco-ops · 1 comment

coilyco-ops commented

2026-07-29 04:39:41 +00:00

Member

Outcome

Generalize seats from harness-shaped identities into stable profile-owned
identity and evaluation contexts.

Parent: #114
Run contract: #115

Preferred schema

seat "chatbot-sonnet-low" name="example content guide" pronouns="they" channel="chatbot" tier="sonnet-low"

The first argument is the stable seat key.
channel and tier are optional descriptive metadata.
The seat selects identity and evaluation context only.
The seat never grants authority, selects credentials, routes a request, or
chooses a live model.
Every person-dependent command selects the stable key through --seat.

Compatibility

Legacy agent "<harness>" entries remain readable throughout v1.x.
The compatibility adapter maps the legacy harness value to both seat key and
channel without changing current generated behavior.
No engine behavior depends on canonical harness or seat names.

Dependencies

#111 supplies the profile and library graph.

Acceptance

Tests prove arbitrary non-harness seat keys and they pronouns.
Compose, evaluation, overlay, roster, catalogue, palette, snapshot, describe,
and verification consume the generalized seat.
Seat metadata appears in deterministic text and structured outputs.
Tests prove that seat selection changes no executable authority.
ward exec test and ward exec smoke pass.
The implementation lands on canonical main and closes this issue.

## Outcome Generalize seats from harness-shaped identities into stable profile-owned identity and evaluation contexts. Parent: #114 Run contract: #115 ## Preferred schema ```kdl seat "chatbot-sonnet-low" name="example content guide" pronouns="they" channel="chatbot" tier="sonnet-low" ``` * The first argument is the stable seat key. * `channel` and `tier` are optional descriptive metadata. * The seat selects identity and evaluation context only. * The seat never grants authority, selects credentials, routes a request, or chooses a live model. * Every person-dependent command selects the stable key through `--seat`. ## Compatibility * Legacy `agent "<harness>"` entries remain readable throughout v1.x. * The compatibility adapter maps the legacy harness value to both seat key and channel without changing current generated behavior. * No engine behavior depends on canonical harness or seat names. ## Dependencies * #111 supplies the profile and library graph. ## Acceptance * Tests prove arbitrary non-harness seat keys and `they` pronouns. * Compose, evaluation, overlay, roster, catalogue, palette, snapshot, describe, and verification consume the generalized seat. * Seat metadata appears in deterministic text and structured outputs. * Tests prove that seat selection changes no executable authority. * `ward exec test` and `ward exec smoke` pass. * The implementation lands on canonical `main` and closes this issue.

coilyco-ops referenced this issue

2026-07-29 04:41:07 +00:00

Adapt persona system to true use case #114

coilyco-ops referenced this issue

2026-07-29 04:41:08 +00:00

Run contract: complete the Agent Compose v1.x issue suite #115

coilyco-ops added the

priority

autonomy

headless

labels

2026-07-29 04:42:03 +00:00

coilyco-ops commented

2026-07-29 05:24:47 +00:00

Author

Member

Landed on canonical main through daa9996, completing the generalized-seat slice begun by earlier commits. Stable seat keys, optional channel and tier metadata, and legacy agent compatibility flow through compose, evaluation, overlay, roster, catalogues, palette, snapshots, describe, and verification. The external fixture proves chatbot-sonnet-low and they pronouns. A direct test proves changing Engineer seats changes identity only, not role skill, doctrine, personalities, protocol, rubric, cases, or authority. ward exec test and ward exec smoke passed.

Landed on canonical main through daa9996, completing the generalized-seat slice begun by earlier commits. Stable seat keys, optional channel and tier metadata, and legacy agent compatibility flow through compose, evaluation, overlay, roster, catalogues, palette, snapshots, describe, and verification. The external fixture proves chatbot-sonnet-low and they pronouns. A direct test proves changing Engineer seats changes identity only, not role skill, doctrine, personalities, protocol, rubric, cases, or authority. ward exec test and ward exec smoke passed.

coilyco-ops closed this issue

2026-07-29 05:24:52 +00:00

coilyco-ops referenced this issue

2026-07-29 06:30:33 +00:00

Run contract: complete the Agent Compose v1.x issue suite #115