hk-local-run-spec

Name: hk-local-run-spec
Author: deepklarity

84stars12forks

How to install hk-local-run-spec?

deepklarity/harness-kit/hk-local-run-spec

$npx -y skills add deepklarity/harness-kit --skill hk-local-run-spec

Installs into the current project.

Summary

Plan and execute an odin spec. Handles working directory, auth, nested-session detection, and post-run diagnostics. Use whenever the user wants to run a spec, test a spec, or do a smoke test. Triggers on: 'run this spec', 'test this spec', 'odin plan', 'smoke test', or /hk-local-

Files · 1

View on GitHub

SKILL.md

1	# /hk-local-run-spec — Spec Runner
2
3	Run the full odin spec workflow: plan, review, execute. Handles the gotchas so you don't have to remember them.
4
5	## Usage
6
7	```
8	/hk-local-run-spec ../sample_specs/poem_spec.md
9	/hk-local-run-spec ../sample_specs/poem_spec.md --quick
10	/hk-local-run-spec ../sample_specs/poem_spec.md --mock
11	/hk-local-run-spec --exec-only a1b2
12	```
13
14	## Arguments
15
16	Parse `$ARGUMENTS` to extract:
17	- spec_path (required unless `--exec-only`): path to the spec markdown file
18	- --quick: use `--quiet` mode (non-interactive, spinner only)
19	- --mock: pass `--mock` to exec (no backend writes)
20	- --exec-only <task_id>: skip planning, just execute a specific task
21
22	## The critical gotcha
23
24	Odin plan/exec invokes `claude -p` as a subprocess. It cannot run from inside another Claude Code session. Nested Claude Code calls fail silently or hang.
25
26	This skill detects the nested session and provides the user with copy-paste commands instead of trying to run them directly.
27
28	## Execution
29
30	### Step 1: Detect environment
31
32	Check if we're inside a Claude Code session:
33
34	```bash
35	# Claude Code sets this env var when running
36	echo "${CLAUDE_CODE_ENTRYPOINT:-not_set}"
37	```
38
39	If we're inside Claude Code (the var is set or we detect we're in an agent context), we cannot run odin plan/exec directly. Instead, provide copy-paste commands.
40
41	### Step 2: Resolve paths
42
43	```bash
44	REPO_ROOT=$(git rev-parse --show-toplevel)
45	WORK_DIR="$REPO_ROOT/odin/temp_test_dir"
46	```
47
48	Check that `temp_test_dir/` exists and has a `.env`:
49
50	```bash
51	ls "$WORK_DIR/.env" 2>/dev/null
52	```
53
54	If missing, tell the user:
55	```
56	temp_test_dir/ not found or missing .env.
57	Create it: cd odin && mkdir -p temp_test_dir && cp .env.example temp_test_dir/.env
58	Then edit temp_test_dir/.env with your ODIN_ADMIN_USER and ODIN_ADMIN_PASSWORD.
59	```
60
61	### Step 3: Generate commands
62
63	Build the commands based on arguments:
64
65	Plan phase (skip if `--exec-only`):
66	```bash
67	cd "$WORK_DIR" && odin plan <spec_path> --quiet
68	```
69
70	If `--quick` was passed, use `--quiet` (which implies `--auto`). Otherwise default to `--auto` for non-interactive execution.
71
72	Review phase (skip if `--exec-only`):
73	```bash
74	cd "$WORK_DIR" && odin specs
75	cd "$WORK_DIR" && odin status
76	```
77
78	Exec phase:
79	```bash
80	cd "$WORK_DIR" && odin exec <task_id> [--mock]
81	```
82
83	### Step 4: Output
84
85	Since we're almost always inside Claude Code (that's where this skill runs), output the commands for the user to copy-paste into a regular terminal:
86
87	```
88	## Spec Run Commands
89
90	Run these from a regular terminal (not inside Claude Code):
91
92	### 1. Plan
93	cd <WORK_DIR> && odin plan <spec_path> --auto
94
95	### 2. Review
96	cd <WORK_DIR> && odin specs
97	cd <WORK_DIR> && odin status
98
99	### 3. Execute (run for each task)
100	cd <WORK_DIR> && odin exec <task_id> [--mock]
101
102	### 4. Post-run diagnostic
103	cd <REPO_ROOT>/taskit/taskit-backend && python testing_tools/spec_trace.py <spec_id> --brief
104	```
105
106	If `--exec-only` was passed, skip the plan/review sections and only show the exec + diagnostic commands.
107
108	### Step 5: Post-run diagnostic hint
109
110	Always remind the user to run the diagnostic after execution:
111
112	```
113	After execution completes, check results with:
114	/hk-local-diagnose spec <spec_id> --brief
115	```
116
117	This connects the two skills into a natural workflow.

People who liked this also viewed

6 picks

Type

hk-local-run-spec

84stars12forks

View on GitHub

How to install hk-local-run-spec?

deepklarity/harness-kit/hk-local-run-spec

$npx -y skills add deepklarity/harness-kit --skill hk-local-run-spec

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# /hk-local-run-spec — Spec Runner
2
3	Run the full odin spec workflow: plan, review, execute. Handles the gotchas so you don't have to remember them.
4
5	## Usage
6
7	```
8	/hk-local-run-spec ../sample_specs/poem_spec.md
9	/hk-local-run-spec ../sample_specs/poem_spec.md --quick
10	/hk-local-run-spec ../sample_specs/poem_spec.md --mock
11	/hk-local-run-spec --exec-only a1b2
12	```
13
14	## Arguments
15
16	Parse `$ARGUMENTS` to extract:
17	- spec_path (required unless `--exec-only`): path to the spec markdown file
18	- --quick: use `--quiet` mode (non-interactive, spinner only)
19	- --mock: pass `--mock` to exec (no backend writes)
20	- --exec-only <task_id>: skip planning, just execute a specific task
21
22	## The critical gotcha
23
24	Odin plan/exec invokes `claude -p` as a subprocess. It cannot run from inside another Claude Code session. Nested Claude Code calls fail silently or hang.
25
26	This skill detects the nested session and provides the user with copy-paste commands instead of trying to run them directly.
27
28	## Execution
29
30	### Step 1: Detect environment
31
32	Check if we're inside a Claude Code session:
33
34	```bash
35	# Claude Code sets this env var when running
36	echo "${CLAUDE_CODE_ENTRYPOINT:-not_set}"
37	```
38
39	If we're inside Claude Code (the var is set or we detect we're in an agent context), we cannot run odin plan/exec directly. Instead, provide copy-paste commands.
40
41	### Step 2: Resolve paths
42
43	```bash
44	REPO_ROOT=$(git rev-parse --show-toplevel)
45	WORK_DIR="$REPO_ROOT/odin/temp_test_dir"
46	```
47
48	Check that `temp_test_dir/` exists and has a `.env`:
49
50	```bash
51	ls "$WORK_DIR/.env" 2>/dev/null
52	```
53
54	If missing, tell the user:
55	```
56	temp_test_dir/ not found or missing .env.
57	Create it: cd odin && mkdir -p temp_test_dir && cp .env.example temp_test_dir/.env
58	Then edit temp_test_dir/.env with your ODIN_ADMIN_USER and ODIN_ADMIN_PASSWORD.
59	```
60
61	### Step 3: Generate commands
62
63	Build the commands based on arguments:
64
65	Plan phase (skip if `--exec-only`):
66	```bash
67	cd "$WORK_DIR" && odin plan <spec_path> --quiet
68	```
69
70	If `--quick` was passed, use `--quiet` (which implies `--auto`). Otherwise default to `--auto` for non-interactive execution.
71
72	Review phase (skip if `--exec-only`):
73	```bash
74	cd "$WORK_DIR" && odin specs
75	cd "$WORK_DIR" && odin status
76	```
77
78	Exec phase:
79	```bash
80	cd "$WORK_DIR" && odin exec <task_id> [--mock]
81	```
82
83	### Step 4: Output
84
85	Since we're almost always inside Claude Code (that's where this skill runs), output the commands for the user to copy-paste into a regular terminal:
86
87	```
88	## Spec Run Commands
89
90	Run these from a regular terminal (not inside Claude Code):
91
92	### 1. Plan
93	cd <WORK_DIR> && odin plan <spec_path> --auto
94
95	### 2. Review
96	cd <WORK_DIR> && odin specs
97	cd <WORK_DIR> && odin status
98
99	### 3. Execute (run for each task)
100	cd <WORK_DIR> && odin exec <task_id> [--mock]
101
102	### 4. Post-run diagnostic
103	cd <REPO_ROOT>/taskit/taskit-backend && python testing_tools/spec_trace.py <spec_id> --brief
104	```
105
106	If `--exec-only` was passed, skip the plan/review sections and only show the exec + diagnostic commands.
107
108	### Step 5: Post-run diagnostic hint
109
110	Always remind the user to run the diagnostic after execution:
111
112	```
113	After execution completes, check results with:
114	/hk-local-diagnose spec <spec_id> --brief
115	```
116
117	This connects the two skills into a natural workflow.

People who liked this also viewed

6 picks

Type