Skip to content

Commit 0b97b8e

Browse files
nhortonnhorton
committed
Add queue check to instructions
1 parent ead6ad6 commit 0b97b8e

5 files changed

Lines changed: 51 additions & 50 deletions

File tree

‎.claude/skills/manual_tests.run_fire_tests/SKILL.md‎

Lines changed: 17 additions & 14 deletions
Original file line numberDiff line numberDiff line change
@@ -83,13 +83,16 @@ For EACH test below, follow this cycle:
8383

8484
1. **Launch a sub-agent** using the Task tool (use a fast model like haiku)
8585
2. **Wait for the sub-agent to complete**
86-
3. **Observe whether the hook fired automatically** - you should see a blocking prompt or command output
87-
4. **Record the result** - pass if hook fired, fail if it didn't
88-
5. **Revert changes**: `git checkout -- manual_tests/`
89-
6. **Proceed to the next test**
86+
3. **Check the queue** - Run `cat .deepwork/queue.json` to verify a queue entry was added for the rule
87+
4. **Observe whether the hook fired automatically** - you should see a blocking prompt or command output
88+
5. **Record the result** - pass if hook fired AND queue entry exists, fail otherwise
89+
6. **Revert changes**: `git checkout -- manual_tests/`
90+
7. **Proceed to the next test**
9091

9192
**IMPORTANT**: Only launch ONE sub-agent at a time. Wait for it to complete and revert before launching the next.
9293

94+
**IMPORTANT**: Always check the queue file before reverting to verify the rule was queued.
95+
9396
### Test Cases (run serially)
9497

9598
**Test 1: Trigger/Safety**
@@ -128,16 +131,16 @@ For EACH test below, follow this cycle:
128131

129132
Record the result after each test:
130133

131-
| Test Case | Should Fire | Hook Fired? | Result |
132-
|-----------|-------------|:-----------:|:------:|
133-
| Trigger/Safety | Edit .py only | | |
134-
| Set Mode | Edit _source.py only | | |
135-
| Pair Mode | Edit _trigger.py only | | |
136-
| Command Action | Edit .txt | | |
137-
| Multi Safety | Edit .py only | | |
138-
| Infinite Block Prompt | Edit .py (no promise) | | |
139-
| Infinite Block Command | Edit .py (no promise) | | |
140-
| Created Mode | Create NEW .yml | | |
134+
| Test Case | Should Fire | Queue Entry? | Hook Fired? | Result |
135+
|-----------|-------------|:------------:|:-----------:|:------:|
136+
| Trigger/Safety | Edit .py only | | | |
137+
| Set Mode | Edit _source.py only | | | |
138+
| Pair Mode | Edit _trigger.py only | | | |
139+
| Command Action | Edit .txt | | | |
140+
| Multi Safety | Edit .py only | | | |
141+
| Infinite Block Prompt | Edit .py (no promise) | | | |
142+
| Infinite Block Command | Edit .py (no promise) | | | |
143+
| Created Mode | Create NEW .yml | | | |
141144

142145
## Quality Criteria
143146

‎.claude/skills/manual_tests.run_not_fire_tests/SKILL.md‎

Lines changed: 0 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -68,10 +68,6 @@ Run all 8 "should NOT fire" tests in **parallel** sub-agents, then verify no blo
6868

6969
Use the Task tool to spawn **ALL of the following sub-agents in a SINGLE message** (parallel execution). Each sub-agent should use a fast model like haiku.
7070

71-
For each test, the sub-agent must:
72-
- Edit BOTH the trigger file AND the safety file
73-
- This satisfies the rule's safety condition, so the rule should NOT fire
74-
7571
**Sub-agent prompts (launch all 8 in parallel):**
7672

7773
a. **Trigger/Safety test** - "Edit `manual_tests/test_trigger_safety_mode/feature.py` to add a comment, AND edit `manual_tests/test_trigger_safety_mode/feature_doc.md` to add a note. Both files must be edited so the rule does NOT fire."

‎.deepwork/jobs/manual_tests/steps/run_fire_tests.md‎

Lines changed: 17 additions & 14 deletions
Original file line numberDiff line numberDiff line change
@@ -36,13 +36,16 @@ For EACH test below, follow this cycle:
3636

3737
1. **Launch a sub-agent** using the Task tool (use a fast model like haiku)
3838
2. **Wait for the sub-agent to complete**
39-
3. **Observe whether the hook fired automatically** - you should see a blocking prompt or command output
40-
4. **Record the result** - pass if hook fired, fail if it didn't
41-
5. **Revert changes**: `git checkout -- manual_tests/`
42-
6. **Proceed to the next test**
39+
3. **Check the queue** - Run `cat .deepwork/queue.json` to verify a queue entry was added for the rule
40+
4. **Observe whether the hook fired automatically** - you should see a blocking prompt or command output
41+
5. **Record the result** - pass if hook fired AND queue entry exists, fail otherwise
42+
6. **Revert changes**: `git checkout -- manual_tests/`
43+
7. **Proceed to the next test**
4344

4445
**IMPORTANT**: Only launch ONE sub-agent at a time. Wait for it to complete and revert before launching the next.
4546

47+
**IMPORTANT**: Always check the queue file before reverting to verify the rule was queued.
48+
4649
### Test Cases (run serially)
4750

4851
**Test 1: Trigger/Safety**
@@ -81,16 +84,16 @@ For EACH test below, follow this cycle:
8184

8285
Record the result after each test:
8386

84-
| Test Case | Should Fire | Hook Fired? | Result |
85-
|-----------|-------------|:-----------:|:------:|
86-
| Trigger/Safety | Edit .py only | | |
87-
| Set Mode | Edit _source.py only | | |
88-
| Pair Mode | Edit _trigger.py only | | |
89-
| Command Action | Edit .txt | | |
90-
| Multi Safety | Edit .py only | | |
91-
| Infinite Block Prompt | Edit .py (no promise) | | |
92-
| Infinite Block Command | Edit .py (no promise) | | |
93-
| Created Mode | Create NEW .yml | | |
87+
| Test Case | Should Fire | Queue Entry? | Hook Fired? | Result |
88+
|-----------|-------------|:------------:|:-----------:|:------:|
89+
| Trigger/Safety | Edit .py only | | | |
90+
| Set Mode | Edit _source.py only | | | |
91+
| Pair Mode | Edit _trigger.py only | | | |
92+
| Command Action | Edit .txt | | | |
93+
| Multi Safety | Edit .py only | | | |
94+
| Infinite Block Prompt | Edit .py (no promise) | | | |
95+
| Infinite Block Command | Edit .py (no promise) | | | |
96+
| Created Mode | Create NEW .yml | | | |
9497

9598
## Quality Criteria
9699

‎.gemini/skills/manual_tests/run_fire_tests.toml‎

Lines changed: 17 additions & 14 deletions
Original file line numberDiff line numberDiff line change
@@ -60,13 +60,16 @@ For EACH test below, follow this cycle:
6060
6161
1. **Launch a sub-agent** using the Task tool (use a fast model like haiku)
6262
2. **Wait for the sub-agent to complete**
63-
3. **Observe whether the hook fired automatically** - you should see a blocking prompt or command output
64-
4. **Record the result** - pass if hook fired, fail if it didn't
65-
5. **Revert changes**: `git checkout -- manual_tests/`
66-
6. **Proceed to the next test**
63+
3. **Check the queue** - Run `cat .deepwork/queue.json` to verify a queue entry was added for the rule
64+
4. **Observe whether the hook fired automatically** - you should see a blocking prompt or command output
65+
5. **Record the result** - pass if hook fired AND queue entry exists, fail otherwise
66+
6. **Revert changes**: `git checkout -- manual_tests/`
67+
7. **Proceed to the next test**
6768
6869
**IMPORTANT**: Only launch ONE sub-agent at a time. Wait for it to complete and revert before launching the next.
6970
71+
**IMPORTANT**: Always check the queue file before reverting to verify the rule was queued.
72+
7073
### Test Cases (run serially)
7174
7275
**Test 1: Trigger/Safety**
@@ -105,16 +108,16 @@ For EACH test below, follow this cycle:
105108
106109
Record the result after each test:
107110
108-
| Test Case | Should Fire | Hook Fired? | Result |
109-
|-----------|-------------|:-----------:|:------:|
110-
| Trigger/Safety | Edit .py only | | |
111-
| Set Mode | Edit _source.py only | | |
112-
| Pair Mode | Edit _trigger.py only | | |
113-
| Command Action | Edit .txt | | |
114-
| Multi Safety | Edit .py only | | |
115-
| Infinite Block Prompt | Edit .py (no promise) | | |
116-
| Infinite Block Command | Edit .py (no promise) | | |
117-
| Created Mode | Create NEW .yml | | |
111+
| Test Case | Should Fire | Queue Entry? | Hook Fired? | Result |
112+
|-----------|-------------|:------------:|:-----------:|:------:|
113+
| Trigger/Safety | Edit .py only | | | |
114+
| Set Mode | Edit _source.py only | | | |
115+
| Pair Mode | Edit _trigger.py only | | | |
116+
| Command Action | Edit .txt | | | |
117+
| Multi Safety | Edit .py only | | | |
118+
| Infinite Block Prompt | Edit .py (no promise) | | | |
119+
| Infinite Block Command | Edit .py (no promise) | | | |
120+
| Created Mode | Create NEW .yml | | | |
118121
119122
## Quality Criteria
120123

‎.gemini/skills/manual_tests/run_not_fire_tests.toml‎

Lines changed: 0 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -46,10 +46,6 @@ Run all 8 "should NOT fire" tests in **parallel** sub-agents, then verify no blo
4646
4747
Use the Task tool to spawn **ALL of the following sub-agents in a SINGLE message** (parallel execution). Each sub-agent should use a fast model like haiku.
4848
49-
For each test, the sub-agent must:
50-
- Edit BOTH the trigger file AND the safety file
51-
- This satisfies the rule's safety condition, so the rule should NOT fire
52-
5349
**Sub-agent prompts (launch all 8 in parallel):**
5450
5551
a. **Trigger/Safety test** - "Edit `manual_tests/test_trigger_safety_mode/feature.py` to add a comment, AND edit `manual_tests/test_trigger_safety_mode/feature_doc.md` to add a note. Both files must be edited so the rule does NOT fire."

0 commit comments

Comments
 (0)