UNPKG

@namastexlabs/speak

Version:

Open source voice dictation for everyone

github.com/namastexlabs/speak

namastexlabs/speak

21 lines (15 loc) • 1.31 kB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
--- name: Evidence & Storage Conventions description: Declare artifact locations and maintain evidence checklists --- # Evidence & Storage Conventions **Last Updated:** !`date -u +"%Y-%m-%d %H:%M:%S UTC"` ## Evidence Requirements - Wishes must declare where artefacts live; there is no default `qa/` directory. Capture metrics inline in the wish (e.g., tables under a **Metrics** section) or in clearly named companion files. - Every wish must complete the **Evidence Checklist** block before implementation begins, spelling out validation commands, artefact locations, and approval checkpoints. - External tracker IDs live in the wish markdown (for example a **Tracking** section with `Forge task: FORGE-123`). - Background agent outputs are summarised in the wish context ledger; raw logs can be viewed with `mcp__genie__view` with sessionId parameter. ## Testing & Evaluation - Evaluation tooling is optional. If a project adds its own evaluator agent, the review or plan workflow can reference it; otherwise, evaluation steps default to manual validation. - Typical metrics: `{{METRICS}}` such as latency or quality. Domain-specific metrics should be added per project in the wish/forge plan. - Validation hooks should be captured in wishes/forge plans (e.g., `pnpm test`, `cargo test`, metrics scripts).