multi-forge 0.2.0__py3-none-any.whl

forge/_extensions/skills/qa/resources/checklist/8-status-line.md

@@ -0,0 +1,174 @@
+<!-- prereq: 0.3, 2.1, 5.1 -->
+
+## 8. Status Line
+
+### 8.1 Direct Invocation
+
+<!-- human:confirm -->
+
+This is a rendered status-line smoke test. It does not call Claude or an LLM; it feeds a synthetic Claude Code
+`statusLine` JSON payload into `forge status-line` and asks you to review the terminal-facing output.
+
+Expected visible shape, with colors/spaces rendered by the terminal:
+
+```text
+${FORGE_TEST_REPO} (main) | test-session-1
+[Opus 4.6] -------- 6%/200K | 3m | +12/-3 | in:28.0K out:17.5K
+```
+
+The output may wrap to two physical terminal lines. A proxy template/tier prefix is expected only when
+`ANTHROPIC_BASE_URL` points at a live or registered Forge proxy; if `test-session-1` was started without a proxy, no
+proxy segment is expected here.
+
+```bash
+cd $FORGE_TEST_REPO
+
+# Mirror Claude Code's statusLine JSON contract and the Forge launch env.
+BASE_URL=$(jq -r '.intent.proxy.base_url // empty' .forge/sessions/test-session-1/forge.session.json)
+mkdir -p .forge/walkthrough
+cat > .forge/walkthrough/status-line-transcript.jsonl <<EOF
+{"requestId":"req-001","message":{"role":"user","content":[{"type":"text","text":"Read the config file."}]}}
+{"requestId":"req-001","message":{"role":"assistant","content":[{"type":"text","text":"I'll inspect it."},{"type":"tool_use","id":"tool-001","name":"Read","input":{"file_path":"${FORGE_TEST_REPO}/config.yaml"}}]}}
+{"requestId":"req-001","message":{"role":"user","content":[{"type":"tool_result","tool_use_id":"tool-001","content":"timeout: 10"}]}}
+{"requestId":"req-002","message":{"role":"user","content":[{"type":"text","text":"Update the timeout and run tests."}]}}
+{"requestId":"req-002","message":{"role":"assistant","content":[{"type":"tool_use","id":"tool-002","name":"Edit","input":{"file_path":"${FORGE_TEST_REPO}/config.yaml"}},{"type":"tool_use","id":"tool-003","name":"Bash","input":{"command":"uv run pytest"}}]}}
+EOF
+STATUS_INPUT=$(jq -nc \
+  --arg cwd "$FORGE_TEST_REPO" \
+  --arg transcript "$FORGE_TEST_REPO/.forge/walkthrough/status-line-transcript.jsonl" \
+  '{
+    workspace: {current_dir: $cwd},
+    model: {display_name: "Opus 4.6"},
+    transcript_path: $transcript,
+    context_window: {
+      context_window_size: 200000,
+      used_percentage: 6,
+      total_input_tokens: 28000,
+      total_output_tokens: 17500,
+      current_usage: {
+        input_tokens: 8500,
+        cache_creation_input_tokens: 2000,
+        cache_read_input_tokens: 1500
+      }
+    },
+    cost: {
+      total_duration_ms: 185000,
+      total_lines_added: 12,
+      total_lines_removed: 3
+    }
+  }')
+
+echo "$STATUS_INPUT" \
+  | FORGE_SESSION=test-session-1 ANTHROPIC_BASE_URL="$BASE_URL" forge status-line
+```
+
+- [ ] Shows compact workspace path, git branch, and `test-session-1`
+- [ ] Shows `[Opus 4.6]` plus context usage `6%/200K` with a visible progress bar
+- [ ] Shows seeded metrics: `3m`, `+12/-3`, `in:28.0K`, and `out:17.5K`
+- [ ] If `ANTHROPIC_BASE_URL` belongs to a created/running proxy, also shows proxy template/tier info
+- [ ] Does not print raw JSON, a Python traceback, or `[Error: ...]`
+- [ ] ANSI/color and non-breaking-space internals are checked in 8.2, not by this rendered review
+
+### 8.2 Verify Display Elements
+
+<!-- human:confirm -->
+
+The status line uses a category-based layout with 5 categories: Where, Who, What, Metrics, State. This step
+intentionally pipes the output through `cat -v`, so the output will look ugly on purpose:
+
+- non-breaking spaces show up as `M-BM-`
+- ANSI escapes show up as `^[[...`
+- colorized line-change segments and dimmed `in:` / `out:` / `cache:` labels still show their raw escapes
+
+Rendered output is covered in 8.1. This step is only checking that the raw escapes and hardened spacing are present.
+
+```bash
+cd $FORGE_TEST_REPO
+
+BASE_URL=$(jq -r '.intent.proxy.base_url // empty' .forge/sessions/test-session-1/forge.session.json)
+mkdir -p .forge/walkthrough
+cat > .forge/walkthrough/status-line-transcript.jsonl <<EOF
+{"requestId":"req-001","message":{"role":"user","content":[{"type":"text","text":"Read the config file."}]}}
+{"requestId":"req-001","message":{"role":"assistant","content":[{"type":"text","text":"I'll inspect it."},{"type":"tool_use","id":"tool-001","name":"Read","input":{"file_path":"${FORGE_TEST_REPO}/config.yaml"}}]}}
+{"requestId":"req-001","message":{"role":"user","content":[{"type":"tool_result","tool_use_id":"tool-001","content":"timeout: 10"}]}}
+{"requestId":"req-002","message":{"role":"user","content":[{"type":"text","text":"Update the timeout and run tests."}]}}
+{"requestId":"req-002","message":{"role":"assistant","content":[{"type":"tool_use","id":"tool-002","name":"Edit","input":{"file_path":"${FORGE_TEST_REPO}/config.yaml"}},{"type":"tool_use","id":"tool-003","name":"Bash","input":{"command":"uv run pytest"}}]}}
+EOF
+STATUS_INPUT=$(jq -nc \
+  --arg cwd "$FORGE_TEST_REPO" \
+  --arg transcript "$FORGE_TEST_REPO/.forge/walkthrough/status-line-transcript.jsonl" \
+  '{
+    workspace: {current_dir: $cwd},
+    model: {display_name: "Opus 4.6 (200k context)"},
+    transcript_path: $transcript,
+    context_window: {
+      context_window_size: 200000,
+      used_percentage: 6,
+      total_input_tokens: 28000,
+      total_output_tokens: 17500,
+      current_usage: {
+        input_tokens: 8500,
+        cache_creation_input_tokens: 2000,
+        cache_read_input_tokens: 1500
+      }
+    },
+    cost: {
+      total_duration_ms: 185000,
+      total_lines_added: 12,
+      total_lines_removed: 3
+    }
+  }')
+
+# Pipe through cat -v to inspect raw ANSI escapes and NBSP rendering.
+# Expected: ugly raw output, not a pretty status line.
+echo "$STATUS_INPUT" \
+  | FORGE_SESSION=test-session-1 ANTHROPIC_BASE_URL="$BASE_URL" forge status-line 2>&1 | cat -v
+# Check for non-breaking spaces (M-BM-), ANSI codes (^[[...),
+# and ASCII indicators/progress-bar text rather than rendered terminal styling.
+```
+
+- [ ] Shows ANSI-colored ASCII segments in raw `cat -v` form
+- [ ] Shows model name (cleaned, without redundant context info)
+- [ ] Uses non-breaking spaces (prevents VSCode trimming)
+- [ ] ANSI reset prefix present
+
+### 8.3 Breadcrumb Display (for resumed sessions)
+
+<!-- human:confirm -->
+
+```bash
+cd $FORGE_TEST_REPO
+
+# Create a disposable derived-looking session so this step does not depend on section 10.
+forge session delete test-session-breadcrumb --force 2>/dev/null || true
+forge session start test-session-breadcrumb --no-launch >/dev/null
+
+cat .forge/sessions/test-session-breadcrumb/forge.session.json \
+  | jq '.confirmed.derivation = {
+      "parent_session": "test-session-1",
+      "parent_transcript": ".forge/artifacts/test-session-1/transcript.jsonl",
+      "inherited_proxy": null,
+      "strategy": "minimal",
+      "depth": 1,
+      "resumed_at": "2026-03-16T00:00:00Z",
+      "lineage": ["test-session-1"]
+    }' > /tmp/test-session-breadcrumb.json && \
+  mv /tmp/test-session-breadcrumb.json .forge/sessions/test-session-breadcrumb/forge.session.json
+
+BASE_URL=$(jq -r '.intent.proxy.base_url // empty' .forge/sessions/test-session-breadcrumb/forge.session.json)
+STATUS_INPUT=$(jq -nc \
+  --arg cwd "$FORGE_TEST_REPO" \
+  '{
+    workspace: {current_dir: $cwd},
+    model: {display_name: "Opus 4.6"}
+  }')
+
+echo "$STATUS_INPUT" \
+  | FORGE_SESSION=test-session-breadcrumb ANTHROPIC_BASE_URL="$BASE_URL" forge status-line 2>/dev/null
+
+forge session delete test-session-breadcrumb --force >/dev/null
+```
+
+- [ ] Shows session lineage breadcrumb (for example `test-session-1 > test-session-breadcrumb`)
+
+---

forge/_extensions/skills/qa/resources/checklist/9-direct-commands.md

@@ -0,0 +1,146 @@
+<!-- prereq: 0.3, 2.1, 5.1 -->
+
+## 9. Direct Commands (% commands)
+
+### 9.1 Test %help
+
+<!-- auto -->
+
+```bash
+# Simulate UserPromptSubmit with %help
+echo '{"prompt": "%help"}' | FORGE_SESSION=test-session-1 forge hook user-prompt-submit
+```
+
+- [ ] Returns help text listing available commands
+
+### 9.2 Test %session list
+
+<!-- auto -->
+
+```bash
+echo '{"prompt": "%session list"}' | FORGE_SESSION=test-session-1 forge hook user-prompt-submit
+```
+
+- [ ] Returns session list (similar to CLI)
+
+### 9.3 Test %proxy list
+
+<!-- auto -->
+
+```bash
+echo '{"prompt": "%proxy list"}' | FORGE_SESSION=test-session-1 forge hook user-prompt-submit
+```
+
+- [ ] Returns proxy list (read-only)
+
+### 9.4 Test %guard commands
+
+<!-- auto -->
+
+```bash
+# Guard status
+echo '{"prompt": "%guard status"}' | FORGE_SESSION=test-session-1 forge hook user-prompt-submit
+
+# Guard enable
+echo '{"prompt": "%guard enable --bundle tdd"}' | FORGE_SESSION=test-session-1 forge hook user-prompt-submit
+
+# Guard disable
+echo '{"prompt": "%guard disable"}' | FORGE_SESSION=test-session-1 forge hook user-prompt-submit
+```
+
+- [ ] `%guard status` returns policy state
+- [ ] `%guard enable` enables TDD enforcement
+- [ ] `%guard disable` disables all policy
+
+### 9.5 Test %guard check (Phase 18)
+
+<!-- auto -->
+
+```bash
+cd $FORGE_TEST_REPO
+
+# Create a test file to generate a diff
+echo 'def hello(): pass' > src/test_guard_check.py
+git add src/test_guard_check.py && git commit -m "placeholder"
+echo 'def hello(): return "world"' > src/test_guard_check.py
+
+# Check unstaged changes against TDD bundle (should deny: impl without tests)
+echo '{"prompt": "%guard check --bundle tdd"}' | FORGE_SESSION=test-session-1 forge hook user-prompt-submit
+
+# Check staged changes
+git add src/test_guard_check.py
+echo '{"prompt": "%guard check --bundle tdd --staged"}' | FORGE_SESSION=test-session-1 forge hook user-prompt-submit
+
+# With explicit bundle override
+echo '{"prompt": "%guard check --bundle coding_standards"}' | FORGE_SESSION=test-session-1 forge hook user-prompt-submit
+
+# Clean up the test file (revert the commit + remove the file)
+git reset --hard HEAD~1
+rm -f src/test_guard_check.py
+
+# Ensure no unstaged changes remain (Forge may have modified settings.local.json etc.)
+git checkout -- . 2>/dev/null || true
+
+# Now verify "no changes" path
+echo '{"prompt": "%guard check --bundle tdd"}' | FORGE_SESSION=test-session-1 forge hook user-prompt-submit
+```
+
+- [ ] `%guard check` returns JSON with `passed`, `files_checked`, `bundles` fields
+- [ ] Impl-only file denied by TDD bundle (`passed: false`)
+- [ ] `--staged` flag evaluates staged changes instead of unstaged
+- [ ] `--bundle` override selects specific bundle
+- [ ] No changes returns `"No unstaged changes to check."`
+
+### 9.6 Test %config
+
+<!-- auto -->
+
+```bash
+echo '{"prompt": "%config"}' | FORGE_SESSION=test-session-1 forge hook user-prompt-submit
+```
+
+- [ ] Returns effective runtime config (read-only)
+
+### 9.7 Test %session list --no-incognito
+
+<!-- auto -->
+
+```bash
+echo '{"prompt": "%session list --no-incognito"}' | FORGE_SESSION=test-session-1 forge hook user-prompt-submit
+```
+
+- [ ] Returns a session list that excludes incognito sessions
+
+### 9.8 Test %proxy show
+
+<!-- auto -->
+
+<!-- prereq: 4.2 -->
+
+```bash
+echo '{"prompt": "%proxy show test-proxy-nostart"}' | FORGE_SESSION=test-session-1 forge hook user-prompt-submit
+```
+
+- [ ] Returns proxy details (template, base_url, status) for the requested proxy id
+
+### 9.9 Test %cancel-verification
+
+<!-- auto -->
+
+```bash
+cd $FORGE_TEST_REPO
+
+# Configure verification (completion promise) on the session
+forge session set verification '{"type":"completion_promise","promise":"FORGE_COMPLETED"}' --session test-session-1
+
+# Enable bypass via direct command escape hatch
+echo '{"prompt": "%cancel-verification"}' | FORGE_SESSION=test-session-1 forge hook user-prompt-submit
+
+# Verify override set
+cat .forge/sessions/test-session-1/forge.session.json | jq '.overrides.verification.bypass'
+```
+
+- [ ] `%cancel-verification` returns "bypass enabled" message
+- [ ] Session overrides include `verification.bypass: true`
+
+---

forge/_extensions/skills/qa/resources/checklist.md

@@ -0,0 +1,103 @@
+# Forge QA Checklist
+
+<!-- checklist: index -->
+
+<!-- version: 1.0.19 -->
+
+<!-- test-count: 482 -->
+
+<!-- last-updated: 2026-05-18 -->
+
+<!-- aligned-with: v0.1.0 -->
+
+**Test Repo**: `$FORGE_TEST_REPO`
+
+**Last updated**: 2026-05-18 (add memory CLI and handoff report checks)
+
+---
+
+## Sections
+
+- [0. Enable Forge (New User Flow)](checklist/0-enable.md)
+
+<!-- section: 0 checklist/0-enable.md -->
+
+- [1. Pre-Flight for Extension Tests](checklist/1-preflight.md)
+
+<!-- section: 1 checklist/1-preflight.md -->
+
+- [2. Claude Code Extensions (`forge extension enable`)](checklist/2-extensions.md)
+
+<!-- section: 2 checklist/2-extensions.md -->
+
+- [3. Authentication (`forge authentication`)](checklist/3-auth.md)
+
+<!-- section: 3 checklist/3-auth.md -->
+
+- [4. Proxy Management](checklist/4-proxy.md)
+
+<!-- section: 4 checklist/4-proxy.md -->
+
+- [5. Session Management](checklist/5-session.md)
+
+<!-- section: 5 checklist/5-session.md -->
+
+- [6. Hooks Testing](checklist/6-hooks.md)
+
+<!-- section: 6 checklist/6-hooks.md -->
+
+- [7. Cost Tracking & Spend Caps](checklist/7-costs.md)
+
+<!-- section: 7 checklist/7-costs.md -->
+
+- [8. Status Line](checklist/8-status-line.md)
+
+<!-- section: 8 checklist/8-status-line.md -->
+
+- [9. Direct Commands (% commands)](checklist/9-direct-commands.md)
+
+<!-- section: 9 checklist/9-direct-commands.md -->
+
+- [10. Session Resume (Phase 10 Feature)](checklist/10-resume.md)
+
+<!-- section: 10 checklist/10-resume.md -->
+
+- [11. Runtime Config + Claude Preset (`forge config`, `forge claude preset`)](checklist/11-config.md)
+
+<!-- section: 11 checklist/11-config.md -->
+
+- [12. Search (`forge search`)](checklist/12-search.md)
+
+<!-- section: 12 checklist/12-search.md -->
+
+- [13. Policy/Guard (`forge guard`)](checklist/13-guard.md)
+
+<!-- section: 13 checklist/13-guard.md -->
+
+- [14. Workflow Runners (`forge workflow`)](checklist/14-workflow.md)
+
+<!-- section: 14 checklist/14-workflow.md -->
+
+- [15. Skills (`/forge:review`, `/forge:understand`, `/forge:panel`, `/forge:consensus`)](checklist/15-skills.md)
+
+<!-- section: 15 checklist/15-skills.md -->
+
+- [16. Handoff Agent](checklist/16-handoff.md)
+
+<!-- section: 16 checklist/16-handoff.md -->
+
+- [17. System Info](checklist/17-info.md)
+
+<!-- section: 17 checklist/17-info.md -->
+
+- [18. Uninstallation (Incremental)](checklist/18-disable.md)
+
+<!-- section: 18 checklist/18-disable.md -->
+
+- [19. Complete Uninstallation (setup.sh --uninstall)](checklist/19-uninstall.md)
+
+<!-- section: 19 checklist/19-uninstall.md -->
+
+- [20. Cleanup](checklist/20-cleanup.md)
+
+<!-- section: 20 checklist/20-cleanup.md -->

forge/_extensions/skills/qa/resources/report-template.md

@@ -0,0 +1,62 @@
+# Forge QA Report
+
+| Field                 | Value                                                   |
+| --------------------- | ------------------------------------------------------- |
+| **Date**              | YYYY-MM-DD                                              |
+| **Forge Version**     | X.Y.Z (output of `forge --version`)                     |
+| **Container**         | container name (from `start-container.sh`)              |
+| **Checklist Version** | X.Y.Z (from checklist header `<!-- version: ... -->`)   |
+| **Duration**          | HH:MM (from state file started_at to last_updated)      |
+| **Debug Logging**     | Enabled by default in QA; artifacts copied when present |
+
+## Summary
+
+| Category              | Total | Pass  | Fail  | Skip  |
+| --------------------- | ----- | ----- | ----- | ----- |
+| Enable (New User)     | 0     | 0     | 0     | 0     |
+| Pre-Flight            | 0     | 0     | 0     | 0     |
+| Extensions            | 0     | 0     | 0     | 0     |
+| Auth                  | 0     | 0     | 0     | 0     |
+| Proxy                 | 0     | 0     | 0     | 0     |
+| Session               | 0     | 0     | 0     | 0     |
+| Hooks                 | 0     | 0     | 0     | 0     |
+| Status Line           | 0     | 0     | 0     | 0     |
+| Direct Commands       | 0     | 0     | 0     | 0     |
+| Session Resume        | 0     | 0     | 0     | 0     |
+| Runtime Config        | 0     | 0     | 0     | 0     |
+| Search                | 0     | 0     | 0     | 0     |
+| Guard                 | 0     | 0     | 0     | 0     |
+| Workflow Runners      | 0     | 0     | 0     | 0     |
+| Skills                | 0     | 0     | 0     | 0     |
+| Handoff               | 0     | 0     | 0     | 0     |
+| System Info           | 0     | 0     | 0     | 0     |
+| Disable (Incremental) | 0     | 0     | 0     | 0     |
+| Uninstall (Complete)  | 0     | 0     | 0     | 0     |
+| Cleanup               | 0     | 0     | 0     | 0     |
+| **TOTAL**             | **0** | **0** | **0** | **0** |
+
+## Issues Found
+
+| #   | Section | Severity        | Description                                   |
+| --- | ------- | --------------- | --------------------------------------------- |
+| 1   | X.Y     | high/medium/low | Brief description of what failed or was wrong |
+
+If no issues: "No issues found."
+
+## Infrastructure
+
+- **Forge**: version, install method (pip/uv)
+- **Docker**: available/unavailable (docker info output)
+- **Proxies**: count from `forge proxy list`, or "not tested"
+- **Credentials**: auth status from hermetic FORGE_HOME, or "not tested"
+
+## Artifacts
+
+- **step-logs/**: raw command output per checklist step (copied from the mounted QA state dir)
+- **forge-logs/final/**: final Forge debug logs copied from the container at artifact-save time
+- **forge-logs-snapshots/**: pre-clean snapshots captured before any checklist step runs `forge logs --clean`
+- **transcript.jsonl**: copied when the QA session exits (if the transcript claim token is satisfied)
+
+## Notes
+
+Observations from human checkpoint verifications, edge cases noticed, or anything that passed but looked suspicious.