docs(session): focus next session on skill load and code-read continuation
This commit is contained in:
ZenchantLive
parent
7a3db2e1c5
commit
f1dc6b1a86
1 changed files with 45 additions and 53 deletions
|
|
@ -1,51 +1,60 @@
|
||||||
# Next Session: Run Epic `beadboard-v5a` (Holistic v5 Audit)
|
# Next Session: Load Skill, Read Code, Continue v5 Audit
|
||||||
|
|
||||||
## Start Here
|
## Objective
|
||||||
|
|
||||||
|
Resume `beadboard-v5a` from current state.
|
||||||
|
|
||||||
|
First priority is context fidelity:
|
||||||
|
1. load/read the skill,
|
||||||
|
2. read the relevant code and tests,
|
||||||
|
3. continue the audit beads in dependency order.
|
||||||
|
|
||||||
|
---
|
||||||
|
|
||||||
|
## Start Commands
|
||||||
|
|
||||||
```bash
|
```bash
|
||||||
cd /mnt/c/Users/Zenchant/codex/beadboard
|
cd /mnt/c/Users/Zenchant/codex/beadboard
|
||||||
|
git status
|
||||||
bd show beadboard-v5a
|
bd show beadboard-v5a
|
||||||
bd ready
|
bd ready
|
||||||
```
|
```
|
||||||
|
|
||||||
You are not starting feature work first. Execute the audit epic and produce critique + remediation graph.
|
---
|
||||||
|
|
||||||
## Epic and Bead Order
|
## Required Skill Load (First)
|
||||||
|
|
||||||
Epic: `beadboard-v5a` — **[EPIC] BeadBoard Driver v5: Holistic Validation and Critique**
|
Read these before touching beads:
|
||||||
|
|
||||||
Execution DAG (already wired):
|
- `skills/beadboard-driver/SKILL.md`
|
||||||
|
- `skills/beadboard-driver/references/coordination-system.md`
|
||||||
|
- `skills/beadboard-driver/references/command-matrix.md`
|
||||||
|
- `skills/beadboard-driver/references/failure-modes.md`
|
||||||
|
- `skills/beadboard-driver/references/session-lifecycle.md`
|
||||||
|
|
||||||
1. `beadboard-v5a.1` baseline verification run
|
---
|
||||||
2. `beadboard-v5a.2` skill-local contract suite execution
|
|
||||||
3. Parallel audit lane after baseline:
|
|
||||||
- `beadboard-v5a.3` SKILL.md runbook dry-run
|
|
||||||
- `beadboard-v5a.4` reference consistency audit
|
|
||||||
- `beadboard-v5a.5` test coverage gap audit
|
|
||||||
- `beadboard-v5a.6` failure-mode drill audit
|
|
||||||
- `beadboard-v5a.10` frontend visual validation gate (manual)
|
|
||||||
- `beadboard-v5a.11` communication system holistic audit
|
|
||||||
- `beadboard-v5a.12` memory system audit
|
|
||||||
- `beadboard-v5a.13` agent lifecycle/liveness audit
|
|
||||||
- `beadboard-v5a.14` swarm/molecule workflow audit
|
|
||||||
- `beadboard-v5a.15` cold-agent usability audit
|
|
||||||
4. `beadboard-v5a.7` consolidated report (`docs/reviews/YYYY-MM-DD-beadboard-driver-v5-audit.md`)
|
|
||||||
5. `beadboard-v5a.8` remediation epic + bead graph creation
|
|
||||||
6. `beadboard-v5a.9` final go/no-go verdict + handoff update
|
|
||||||
|
|
||||||
## Required Skills For This Session
|
## Required Code Read (Second)
|
||||||
|
|
||||||
Use these skills explicitly:
|
Read these implementation files before making claims:
|
||||||
|
|
||||||
1. `beadboard-driver`
|
- `tools/bb.ts`
|
||||||
2. `verification-before-completion`
|
- `src/cli/beadboard-cli.ts`
|
||||||
3. `systematic-debugging` (for any failing gate/drill)
|
- `src/lib/agent-mail.ts`
|
||||||
4. `writing-skills` (critiquing SKILL.md + reference quality)
|
- `skills/beadboard-driver/scripts/bb-mail-shim.mjs`
|
||||||
5. `writing-plans` (when translating findings into remediation plan)
|
- `skills/beadboard-driver/scripts/session-preflight.mjs`
|
||||||
|
- `skills/beadboard-driver/scripts/ensure-bb-mail-configured.mjs`
|
||||||
|
- `skills/beadboard-driver/tests/run-tests.mjs`
|
||||||
|
|
||||||
## Required Verification
|
---
|
||||||
|
|
||||||
At minimum, run and capture outputs:
|
## Continue Work
|
||||||
|
|
||||||
|
- Continue `beadboard-v5a` beads in dependency order.
|
||||||
|
- If `beadboard-v5a.1` is still open/in_progress, finish it first.
|
||||||
|
- Record evidence in bead notes before closing each bead.
|
||||||
|
|
||||||
|
Core verification commands:
|
||||||
|
|
||||||
```bash
|
```bash
|
||||||
npm run typecheck
|
npm run typecheck
|
||||||
|
|
@ -54,27 +63,10 @@ npm run test
|
||||||
node skills/beadboard-driver/tests/run-tests.mjs
|
node skills/beadboard-driver/tests/run-tests.mjs
|
||||||
```
|
```
|
||||||
|
|
||||||
Also capture manual visual evidence for frontend/comms validation (`v5a.10`): screenshots + explicit human confirmation note.
|
---
|
||||||
|
|
||||||
## Bead Authoring Rule (When Creating Remediation Beads)
|
## Rules
|
||||||
|
|
||||||
Follow:
|
|
||||||
|
|
||||||
- `docs/protocols/bead-prompting.md`
|
|
||||||
|
|
||||||
Every new remediation bead description must include:
|
|
||||||
|
|
||||||
- `TASK CONTEXT`
|
|
||||||
- `TASK CONTRACT` (Goal, Success Criteria, Scope, Out of Scope)
|
|
||||||
- `IMPLEMENTATION CONSTRAINTS`
|
|
||||||
- `VERIFICATION REQUIREMENTS`
|
|
||||||
|
|
||||||
## Deliverables (Definition of Done)
|
|
||||||
|
|
||||||
Session is done only when all are true:
|
|
||||||
|
|
||||||
1. `beadboard-v5a.1` through `beadboard-v5a.15` are completed per dependency order.
|
|
||||||
2. Consolidated audit report committed under `docs/reviews/`.
|
|
||||||
3. Remediation epic/bead graph created and linked with correct dependencies.
|
|
||||||
4. Go/no-go verdict written with evidence and residual risks.
|
|
||||||
|
|
||||||
|
- Use `--assignee` on in-progress updates.
|
||||||
|
- Do not close beads without fresh evidence in current session.
|
||||||
|
- If you create remediation beads, follow `docs/protocols/bead-prompting.md`.
|
||||||
|
|
|
||||||
Loading…
Add table
Add a link
Reference in a new issue