savant-code 0.0.19 → 0.0.20

package/dev/test-prompts/archive/release-az-test-fid-2026-0728-launch-tracks.md

@@ -1,253 +0,0 @@
-<!-- markdownlint-disable MD013 MD022 MD032 MD060 -->
-# Historical Release A-Z Test — Launch Tracks v0.0.9
-
-> Historical record from 2026-07-28. This prompt is preserved for audit history and is not current certification for v0.0.11.
-
-**Version:** v0.0.9
-**Purpose:** Regression and feature verification for the v0.0.9 launch tracks: Trust & Verification (FID-003), Safety Track (FID-004), Friction Reduction (FID-005), Launch Artifacts (FID-006), and Master Launch Strategy Execution (FID-007).
-
-**Ground Rules:**
-- Run from agent context (idle phase unless noted)
-- Do not require user interaction
-- Report pass/fail and any friction for every test
-- Write the final report to `dev/scratchpad/release-az-test-fid-2026-0728-launch-tracks-report.md`
-
-**Available Tools:** read_files, glob, list_directory, spawn_agents, write_todos, basher, code_searcher
-
----
-
-## Tier 1: Build & Type Safety
-
-### T1.1 — Common workspace typecheck
-- Run `cd common && bun run typecheck`
-- **Expected:** exit code 0, no errors
-
-### T1.2 — Agent-runtime workspace typecheck
-- Run `cd packages/agent-runtime && bun run typecheck`
-- **Expected:** exit code 0, no errors
-
-### T1.3 — SDK workspace typecheck
-- Run `cd sdk && bun run typecheck`
-- **Expected:** exit code 0, no errors
-
-### T1.4 — CLI workspace typecheck
-- Run `cd cli && bun run typecheck`
-- **Expected:** exit code 0, no errors
-
-### T1.5 — llm-providers workspace typecheck
-- Run `cd packages/llm-providers && bun run typecheck`
-- **Expected:** exit code 0, no errors
-
-### T1.6 — ESLint zero warnings on changed areas
-- Run `bun x eslint cli/src/commands/health-command.ts cli/src/utils/ollama-onboarding.ts cli/src/utils/settings.ts cli/src/utils/openrouter-models.ts packages/llm-providers/src/ollama/detect.ts packages/llm-providers/src/ollama/index.ts --max-warnings 0`
-- **Expected:** zero warnings, zero errors
-
-### T1.7 — Version metadata
-- Read `VERSION`
-- Read root `package.json`
-- **Expected:** both report `0.0.9`
-
----
-
-## Tier 2: Trust & Verification Track (FID-2026-0728-003)
-
-### T2.1 — Telemetry defaults to opt-in (off)
-- Read `cli/src/utils/settings.ts`
-- Verify the default value of `adsEnabled` / telemetry setting is `false`
-- **Expected:** new users start with telemetry/ads disabled
-
-### T2.2 — Secret redaction in logger
-- Read `cli/src/utils/logger.ts`
-- Verify `sanitizeSecrets` is defined and called before values are logged
-- **Expected:** values whose keys look like secrets/tokens are redacted
-
-### T2.3 — Config directory uses `.savant-code`
-- Search `common/src`, `cli/src`, `sdk/src`, and `packages/agent-runtime/src` for config directory name references
-- Verify `.savant-code` is used (not `.savant`)
-- **Expected:** all runtime paths use `.savant-code[-env]/`
-
-### T2.4 — Privacy documentation
-- Read `docs/privacy.md`
-- Verify it documents data boundaries, credential storage, network calls, retention, and user controls
-- **Expected:** privacy architecture is documented
-
----
-
-## Tier 3: Safety Track (FID-2026-0728-004)
-
-### T3.1 — Safety registry exists
-- Read `common/src/tools/safety.ts` and `common/src/tools/safety-registry.ts`
-- Verify `safety-registry.ts` exports a registry mapping tool names to safety metadata (level: safe | prompt | unsafe)
-- **Expected:** all registered tools have safety metadata
-
-### T3.2 — `/permissions` slash command registered
-- Read `cli/src/commands/command-registry.ts`
-- Verify `handlePermissionsCommand` is registered with `/permissions`, `/sandbox`, and `/safety` aliases
-- **Expected:** command is accessible via all three slash-command names
-
-### T3.3 — `/permissions` accepts valid modes
-- Read `cli/src/commands/permissions-command.ts`
-- Verify it accepts `safe`, `prompt`, and `unsafe` as valid mode arguments
-- **Expected:** mode is stored and applied to the current session
-
-### T3.4 — `/permissions` rejects invalid mode
-- Read `cli/src/commands/permissions-command.ts`
-- Verify it returns an error message for modes other than safe/prompt/unsafe
-- **Expected:** clear error message listing valid modes
-
-### T3.5 — Sandbox engine denylist
-- Search `packages/agent-runtime/src` for sandbox or safety enforcement logic
-- Verify destructive shell commands (rm -rf, git push --force, etc.) are denied in safe mode
-- **Expected:** denylist blocks destructive operations when in safe mode
-
-### T3.6 --permission-mode CLI flag
-- Search `cli/src` for `permission-mode` or `permissionMode` flag parsing
-- Verify the CLI parses `--permission-mode safe|prompt|unsafe` at startup
-- **Expected:** flag sets initial permission mode before any commands run
-
-### T3.7 — Network gating
-- Read `packages/agent-runtime/src/tools/sandbox/engine.ts`
-- Verify `createDefaultSandboxPolicy` sets `allowNetwork` based on permission mode
-- Verify `evaluateToolCall` denies network tools in `safe` mode, returns `prompt` in `prompt` mode, and allows them in `unsafe` mode
-- **Expected:** network access respects permission mode settings
-
----
-
-## Tier 4: Friction Reduction Track (FID-2026-0728-005)
-
-### T4.1 — Ollama detection utility exists
-- Read `packages/llm-providers/src/ollama/detect.ts`
-- Verify it probes `/api/version` and `/api/tags`, honoring `OLLAMA_HOST`
-- **Expected:** detection returns model list and version when Ollama is running
-
-### T4.2 — Ollama detection tests pass
-- Run `cd packages/llm-providers && bun test src/ollama/__tests__/detect.test.ts`
-- **Expected:** all tests pass
-
-### T4.3 — First-run Ollama onboarding
-- Read `cli/src/utils/ollama-onboarding.ts`
-- Verify it auto-configures direct-provider mode when Ollama is detected
-- Verify it persists the direct-provider choice across sessions
-- **Expected:** first-run users with Ollama are not prompted for provider setup
-
-### T4.4 — Onboarding tests pass
-- Run `cd cli && bun test src/utils/__tests__/ollama-onboarding.test.ts`
-- **Expected:** all tests pass
-
-### T4.5 — `/health` slash command
-- Read `cli/src/commands/health-command.ts`
-- Verify it reports Ollama status, provider mode, model preference, and permission mode
-- Verify it is registered in `cli/src/commands/command-registry.ts`
-- **Expected:** `/health` is accessible
-
----
-
-## Tier 5: Launch Artifacts Track (FID-2026-0728-006)
-
-### T5.1 — docs/launch directory exists
-- List `docs/launch/`
-- Verify all planned files exist: `hn-post.md`, `hn-first-comment.md`, `twitter-thread.md`, `mastodon-thread.md`, `newsletter-pitch.md`, `incident-response.md`
-- **Expected:** all files present
-
-### T5.2 — Landing page static files
-- List `docs/launch/landing/`
-- Verify `index.html` exists
-- **Expected:** static landing page is ready for deployment
-
-### T5.3 — README hero updated
-- Read `README.md`
-- Verify the hero section has a one-sentence pitch, install command, and Ollama setup instructions
-- **Expected:** README is optimized for launch traffic
-
-### T5.4 — Markdown lint on launch docs
-- Run `bun x markdownlint-cli2 docs/launch/*.md --config .markdownlint.json`
-- **Expected:** zero issues
-
----
-
-## Tier 6: Master Launch Strategy (FID-2026-0728-007)
-
-### T6.1 — All child FIDs archived
-- Read `dev/fids/archive/FID-2026-0728-003-launch-trust-verification.md`
-- Read `dev/fids/archive/FID-2026-0728-004-launch-safety-track.md`
-- Read `dev/fids/archive/FID-2026-0728-005-launch-friction-reduction.md`
-- Read `dev/fids/archive/FID-2026-0728-006-launch-artifacts-track.md`
-- Verify all have status `closed` and archived date set
-- **Expected:** all child tracks are closed
-
-### T6.2 — Master FID archived
-- Read `dev/fids/archive/FID-2026-0728-007-launch-strategy-master.md`
-- Verify status is `closed` and Launch Captain is assigned
-- **Expected:** master FID is closed
-
-### T6.3 — CHANGELOG entries
-- Read `CHANGELOG.md`
-- Verify entries for FID-003 through FID-007 are present under v0.0.9
-- **Expected:** CHANGELOG documents all launch tracks
-
----
-
-## Tier 7: Regression Checks
-
-### T7.1 — `/goal` and `/loop` still work
-- Read `cli/src/commands/goal.ts` and `cli/src/commands/loop.ts`
-- Verify handlers are still registered in `cli/src/commands/command-registry.ts`
-- **Expected:** goal/loop commands remain accessible
-
-### T7.2 — `/login` and `/signin` still work
-- Read `cli/src/commands/command-registry.ts`
-- Verify `/login` and `/signin` aliases are registered
-- **Expected:** login command remains accessible
-
-### T7.3 — No stale `.savant` references in production
-- Search `common/src`, `cli/src`, `sdk/src`, and `packages/agent-runtime/src` for `savant` (excluding test files)
-- **Expected:** zero matches in production source
-
----
-
-## Tier 8: CLI Smoke (if tmux available)
-
-### T8.1 — CLI launches
-- If possible, launch the CLI with `bun run src/index.tsx --cwd ..` from `cli/`
-- Verify it starts without crashing
-- **Expected:** prompt appears
-
-### T8.2 — `/health` returns status
-- In the CLI, run `/health`
-- **Expected:** health report is displayed with provider and Ollama status
-
-### T8.3 — `/permissions safe` sets safe mode
-- In the CLI, run `/permissions safe`
-- **Expected:** permission mode updates to `safe`
-
-### T8.4 — Ollama auto-detection on first launch
-- Clear any saved settings and launch the CLI with Ollama running
-- **Expected:** CLI routes to local Ollama without prompting for provider setup
-
----
-
-## Report Format
-
-After all tiers, write `dev/scratchpad/release-az-test-fid-2026-0728-launch-tracks-report.md` with:
-
-1. **Executive Summary** — 3-5 sentences on v0.0.9 launch readiness
-2. **Tier-by-Tier Results** — For each test: Status, Notes, Friction Level (none/low/medium/high)
-3. **Blockers** — Any test that must be fixed before release
-4. **Pre-existing Issues** — Any failures not caused by this feature
-5. **Release Recommendation** — Go / No-Go with justification
-
----
-
-## Summary
-
-| Tier | Name | Tests | Purpose |
-|------|------|-------|---------|
-| 1 | Build & Type Safety | 7 | Does the code compile and pass lint? |
-| 2 | Trust & Verification | 4 | Is the privacy/BYOK positioning hardened? |
-| 3 | Safety Track | 7 | Is the sandbox engine complete and wired? |
-| 4 | Friction Reduction | 5 | Is Ollama onboarding frictionless? |
-| 5 | Launch Artifacts | 4 | Are launch assets present and lint-clean? |
-| 6 | Master Launch Strategy | 3 | Are all launch FIDs closed and documented? |
-| 7 | Regression Checks | 3 | Are previous features still intact? |
-| 8 | CLI Smoke | 4 | Does the feature hold up in the real CLI? |
-| **Total** | | **37** | |

package/dev/test-prompts/archive/release-az-test-fid-2026-0731-pre-launch.md

@@ -1,102 +0,0 @@
-<!-- markdownlint-disable MD013 MD022 MD032 MD060 -->
-# Current Release A–Z Audit — Savant-Code v0.0.11
-
-**Version:** v0.0.11  
-**FID:** FID-2026-0731-003  
-**Purpose:** Fresh pre-launch evidence for the currently published package. Historical v0.0.9 prompts and reports are not certification for this version.
-
-## Ground Rules
-
-- Run from the repository root unless a command changes directory explicitly.
-- Record exact exit codes and concise output for every check.
-- Use `PASS`, `FAIL`, or `DEFERRED`; never convert unavailable interactive/backend checks into `PASS`.
-- Do not publish, upload, promote, advertise, commit, or push.
-- Telemetry/privacy behavior is governed by the closed FID-2026-0731-006 decision and is validated as a bounded control surface by this audit.
-
-## Tier 1 — Version and Package Identity
-
-### T1.1 — Release sources
-Read `VERSION`, root `package.json`, `cli/package.json`, `sdk/package.json`, and `protocol.config.yaml`. Expected: release sources report `0.0.11`; protocol version remains `0.2.0`; internal workspace packages may retain their documented `0.0.1` versions.
-
-### T1.2 — Registry version
-Run `npm view savant-code version`. Expected: `0.0.11`.
-
-### T1.3 — Package dry-runs
-Run:
-
-```text
-npm pack ./cli/release --dry-run
-npm pack ./cli/release-staging --dry-run
-npm pack ./savant-free/cli/release --dry-run
-```
-
-Expected: all supported package contracts exit 0; staging is private/internal-only.
-
-## Tier 2 — Workspace Safety
-
-### T2.1 — Common typecheck
-`cd common && bun run typecheck`
-
-### T2.2 — Agent-runtime typecheck
-`cd packages/agent-runtime && bun run typecheck`
-
-### T2.3 — SDK typecheck
-`cd sdk && bun run typecheck`
-
-### T2.4 — CLI typecheck
-`cd cli && bun run typecheck`
-
-### T2.5 — LLM-provider typecheck
-`cd packages/llm-providers && bun run typecheck`
-
-### T2.6 — Focused release regression tests
-Run the wrapper-safety, proxy HTTP, terminal-reset, and settings tests.
-
-## Tier 3 — Safety and Runtime Claims
-
-### T3.1 — Permission controls
-Verify `/permissions`, `/sandbox`, `/safety`, and `--permission-mode safe|prompt|unsafe` are registered and parsed.
-
-### T3.2 — Sandbox enforcement
-Verify the sandbox registry, destructive-command denylist, and network gating are present in the runtime and covered by tests.
-
-### T3.3 — Onboarding/health
-Verify Ollama detection/onboarding and `/health` sources exist and have focused tests.
-
-### T3.4 — Regression command paths
-Verify `/goal`, `/loop`, `/login`, and `/signin` remain registered.
-
-## Tier 4 — Public Claims and Governance
-
-### T4.1 — FID gate
-Verify the v0.0.11 report does not inherit historical Go decisions and that the master/child FIDs remain the controlling gate.
-
-### T4.2 — Telemetry boundary
-Verify FID-2026-0731-006 remains archived `closed`, its control surface is tested, and no broader promotion claim is inferred.
-
-### T4.3 — Documentation inventory
-Verify current README/landing/launch artifacts are covered by FID-2026-0731-005 and that placeholders/stale claims are not silently certified.
-
-### T4.4 — Lifecycle inventory
-Verify FID-2026-0731-004 evidence and report any remaining historical metadata backlog honestly.
-
-## Tier 5 — Interactive and Environment-Dependent Checks
-
-### T5.1 — CLI launch
-**DEFERRED** unless a configured backend and interactive terminal are available; use tmux when running this check.
-
-### T5.2 — Auth/model selection
-**DEFERRED** unless credentials/backend are available. Do not fabricate a successful backend call.
-
-### T5.3 — Cross-platform packaging
-**DEFERRED** for platforms not exercised in this Windows environment.
-
-## Report Contract
-
-Write `dev/scratchpad/release-az-test-fid-2026-0731-pre-launch-report.md` with:
-
-1. Exact command/file evidence for every tier.
-2. PASS/FAIL/DEFERRED status and friction.
-3. Blockers and pre-existing issues.
-4. Version-source classification.
-5. A fresh Go/No-Go recommendation that treats unresolved release, FID, docs, telemetry, or deferred critical evidence as a No-Go for promotion.