dual-axis-skill-reviewer

Name: dual-axis-skill-reviewer
Author: mphinance

15stars3forks

How to install dual-axis-skill-reviewer?

mphinance/alpha-skills/dual-axis-skill-reviewer

$npx -y skills add mphinance/alpha-skills --skill dual-axis-skill-reviewer

Installs into the current project.

Summary

Review skills in any project using a dual-axis method: (1) deterministic code-based checks (structure, scripts, tests, execution safety) and (2) LLM deep review findings. Use when you need reproducible quality scoring for skills/*/SKILL.md, want to gate merges with a score thre

Files · 1

View on GitHub

SKILL.md

1	# Dual Axis Skill Reviewer
2
3	Run the dual-axis reviewer script and save reports to `reports/`.
4
5	The script supports:
6	- Random or fixed skill selection
7	- Auto-axis scoring with optional test execution
8	- LLM prompt generation
9	- LLM JSON review merge with weighted final score
10	- Cross-project review via `--project-root`
11
12	## When to Use
13
14	- Need reproducible scoring for one skill in `skills/*/SKILL.md`.
15	- Need improvement items when final score is below 90.
16	- Need both deterministic checks and qualitative LLM code/content review.
17	- Need to review skills in a different project from the command line.
18
19	## Prerequisites
20
21	- Python 3.9+
22	- `uv` (recommended — auto-resolves `pyyaml` dependency via inline metadata)
23	- For tests: `uv sync --extra dev` or equivalent in the target project
24	- For LLM-axis merge: JSON file that follows the LLM review schema (see Resources)
25
26	## Workflow
27
28	Determine the correct script path based on your context:
29
30	- Same project: `skills/dual-axis-skill-reviewer/scripts/run_dual_axis_review.py`
31	- Global install: `~/.claude/skills/dual-axis-skill-reviewer/scripts/run_dual_axis_review.py`
32
33	The examples below use `REVIEWER` as a placeholder. Set it once:
34
35	```bash
36	# If reviewing from the same project:
37	REVIEWER=skills/dual-axis-skill-reviewer/scripts/run_dual_axis_review.py
38
39	# If reviewing another project (global install):
40	REVIEWER=~/.claude/skills/dual-axis-skill-reviewer/scripts/run_dual_axis_review.py
41	```
42
43	### Step 1: Run Auto Axis + Generate LLM Prompt
44
45	```bash
46	uv run "$REVIEWER" \
47	--project-root . \
48	--emit-llm-prompt \
49	--output-dir reports/
50	```
51
52	When reviewing a different project, point `--project-root` to it:
53
54	```bash
55	uv run "$REVIEWER" \
56	--project-root /path/to/other/project \
57	--emit-llm-prompt \
58	--output-dir reports/
59	```
60
61	### Step 2: Run LLM Review
62	- Use the generated prompt file in `reports/skill_review_prompt_<skill>_<timestamp>.md`.
63	- Ask the LLM to return strict JSON output.
64	- When running inside Claude Code, let Claude act as orchestrator: read the generated prompt, produce the LLM review JSON, and save it for the merge step.
65
66	### Step 3: Merge Auto + LLM Axes
67
68	```bash
69	uv run "$REVIEWER" \
70	--project-root . \
71	--skill <skill-name> \
72	--llm-review-json <path-to-llm-review.json> \
73	--auto-weight 0.5 \
74	--llm-weight 0.5 \
75	--output-dir reports/
76	```
77
78	### Step 4: Optional Controls
79
80	- Fix selection for reproducibility: `--skill <name>` or `--seed <int>`
81	- Review all skills at once: `--all`
82	- Skip tests for quick triage: `--skip-tests`
83	- Change report location: `--output-dir <dir>`
84	- Increase `--auto-weight` for stricter deterministic gating.
85	- Increase `--llm-weight` when qualitative/code-review depth is prioritized.
86
87	## Output
88
89	- `reports/skill_review_<skill>_<timestamp>.json`
90	- `reports/skill_review_<skill>_<timestamp>.md`
91	- `reports/skill_review_prompt_<skill>_<timestamp>.md` (when `--emit-llm-prompt` is enabled)
92
93	## Installation (Global)
94
95	To use this skill from any project, symlink it into `~/.claude/skills/`:
96
97	```bash
98	ln -sfn /path/to/claude-trading-skills/skills/dual-axis-skill-reviewer \
99	~/.claude/skills/dual-axis-skill-reviewer
100	```
101
102	After this, Claude Code will discover the skill in all projects, and the script is accessible at `~/.claude/skills/dual-axis-skill-reviewer/scripts/run_dual_axis_review.py`.
103
104	## Resources
105
106	- Auto axis scores metadata, workflow coverage, execution safety, artifact presence, and test health.
107	- Auto axis detects `knowledge_only` skills and adjusts script/test expectations to avoid unfair penalties.
108	- LLM axis scores deep content quality (correctness, risk, missing logic, maintainability).
109	- Final score is weighted average.
110	- If final score is below 90, improvement items are required and listed in the markdown report.
111	- Script: `skills/dual-axis-skill-reviewer/scripts/run_dual_axis_review.py`
112	- LLM schema: `references/llm_review_schema.md`
113	- Rubric detail: `references/scoring_rubric.md`

People who liked this also viewed

6 picks

Type

dual-axis-skill-reviewer

15stars3forks

View on GitHub

How to install dual-axis-skill-reviewer?

mphinance/alpha-skills/dual-axis-skill-reviewer

$npx -y skills add mphinance/alpha-skills --skill dual-axis-skill-reviewer

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# Dual Axis Skill Reviewer
2
3	Run the dual-axis reviewer script and save reports to `reports/`.
4
5	The script supports:
6	- Random or fixed skill selection
7	- Auto-axis scoring with optional test execution
8	- LLM prompt generation
9	- LLM JSON review merge with weighted final score
10	- Cross-project review via `--project-root`
11
12	## When to Use
13
14	- Need reproducible scoring for one skill in `skills/*/SKILL.md`.
15	- Need improvement items when final score is below 90.
16	- Need both deterministic checks and qualitative LLM code/content review.
17	- Need to review skills in a different project from the command line.
18
19	## Prerequisites
20
21	- Python 3.9+
22	- `uv` (recommended — auto-resolves `pyyaml` dependency via inline metadata)
23	- For tests: `uv sync --extra dev` or equivalent in the target project
24	- For LLM-axis merge: JSON file that follows the LLM review schema (see Resources)
25
26	## Workflow
27
28	Determine the correct script path based on your context:
29
30	- Same project: `skills/dual-axis-skill-reviewer/scripts/run_dual_axis_review.py`
31	- Global install: `~/.claude/skills/dual-axis-skill-reviewer/scripts/run_dual_axis_review.py`
32
33	The examples below use `REVIEWER` as a placeholder. Set it once:
34
35	```bash
36	# If reviewing from the same project:
37	REVIEWER=skills/dual-axis-skill-reviewer/scripts/run_dual_axis_review.py
38
39	# If reviewing another project (global install):
40	REVIEWER=~/.claude/skills/dual-axis-skill-reviewer/scripts/run_dual_axis_review.py
41	```
42
43	### Step 1: Run Auto Axis + Generate LLM Prompt
44
45	```bash
46	uv run "$REVIEWER" \
47	--project-root . \
48	--emit-llm-prompt \
49	--output-dir reports/
50	```
51
52	When reviewing a different project, point `--project-root` to it:
53
54	```bash
55	uv run "$REVIEWER" \
56	--project-root /path/to/other/project \
57	--emit-llm-prompt \
58	--output-dir reports/
59	```
60
61	### Step 2: Run LLM Review
62	- Use the generated prompt file in `reports/skill_review_prompt_<skill>_<timestamp>.md`.
63	- Ask the LLM to return strict JSON output.
64	- When running inside Claude Code, let Claude act as orchestrator: read the generated prompt, produce the LLM review JSON, and save it for the merge step.
65
66	### Step 3: Merge Auto + LLM Axes
67
68	```bash
69	uv run "$REVIEWER" \
70	--project-root . \
71	--skill <skill-name> \
72	--llm-review-json <path-to-llm-review.json> \
73	--auto-weight 0.5 \
74	--llm-weight 0.5 \
75	--output-dir reports/
76	```
77
78	### Step 4: Optional Controls
79
80	- Fix selection for reproducibility: `--skill <name>` or `--seed <int>`
81	- Review all skills at once: `--all`
82	- Skip tests for quick triage: `--skip-tests`
83	- Change report location: `--output-dir <dir>`
84	- Increase `--auto-weight` for stricter deterministic gating.
85	- Increase `--llm-weight` when qualitative/code-review depth is prioritized.
86
87	## Output
88
89	- `reports/skill_review_<skill>_<timestamp>.json`
90	- `reports/skill_review_<skill>_<timestamp>.md`
91	- `reports/skill_review_prompt_<skill>_<timestamp>.md` (when `--emit-llm-prompt` is enabled)
92
93	## Installation (Global)
94
95	To use this skill from any project, symlink it into `~/.claude/skills/`:
96
97	```bash
98	ln -sfn /path/to/claude-trading-skills/skills/dual-axis-skill-reviewer \
99	~/.claude/skills/dual-axis-skill-reviewer
100	```
101
102	After this, Claude Code will discover the skill in all projects, and the script is accessible at `~/.claude/skills/dual-axis-skill-reviewer/scripts/run_dual_axis_review.py`.
103
104	## Resources
105
106	- Auto axis scores metadata, workflow coverage, execution safety, artifact presence, and test health.
107	- Auto axis detects `knowledge_only` skills and adjusts script/test expectations to avoid unfair penalties.
108	- LLM axis scores deep content quality (correctness, risk, missing logic, maintainability).
109	- Final score is weighted average.
110	- If final score is below 90, improvement items are required and listed in the markdown report.
111	- Script: `skills/dual-axis-skill-reviewer/scripts/run_dual_axis_review.py`
112	- LLM schema: `references/llm_review_schema.md`
113	- Rubric detail: `references/scoring_rubric.md`

People who liked this also viewed

6 picks

Type