skill-forge-audit

Name: skill-forge-audit
Author: WrathZA

17stars0forks

How to install skill-forge-audit?

wrathza/skillforge/skill-forge-audit

$npx -y skills add WrathZA/skillforge --skill skill-forge-audit

Installs into the current project.

Summary

Batch-evaluate all skills in the repo with skill-forge-judge and render a single consolidated grade report sorted by grade (worst first) so effort is directed correctly. Use when reviewing overall skill quality, finding where to invest improvement effort, or after bulk skill chan

Files · 1

View on GitHub

SKILL.md

1	# Skill Forge Audit
2
3	Batch-run skill-forge-judge across every skill in the repo. One consolidated report — grades, top issues, priority order. Read-only: no fixes applied.
4
5	Requires `skill-forge-judge` to be installed.
6
7	No `references/` — this skill is single-file by design; the whole workflow fits in the body, so the absent directory is intentional, not missing progressive disclosure.
8
9	---
10
11	## Workflow
12
13	### Phase 1 — Discover
14
15	Glob for all SKILL.md files, run from the repo root. Skills live one level deep under `skills/`:
16
17	```
18	pattern: skills/*/SKILL.md
19	```
20
21	Collect the directory name of each match — that's the skill name.
22
23	To detect skipped directories: also Glob `skills//` to list every skill directory. Any directory in the `skills//` result that is not in the `skills/*/SKILL.md` result has no SKILL.md — list it in the Skipped section of the report with reason "No SKILL.md found".
24
25	If zero SKILL.md files found: output "No skills found in this repo." and stop.
26
27	### Phase 2 — Evaluate
28
29	For each discovered skill, in order:
30
31	1. Read the SKILL.md file at its absolute path.
32	2. Invoke `/skill-forge-judge` on it.
33	3. From the judge output, extract and record:
34	- Skill name
35	- Grade (A/B/C/D/F)
36	- Score (X/Y, Z%)
37	- Numbered Improvements: up to 5 items (all if fewer than 5 exist)
38
39	Emit a one-line status per skill as each evaluation completes — do not wait until all are done:
40	```
41	✓ skill-forge-create B (98/120, 82%)
42	✓ skill-forge-hitl C (90/120, 75%)
43	```
44
45	If `/skill-forge-judge` is not installed: output "skill-forge-judge is required — install it before running skill-forge-audit." and stop.
46
47	If N > 10: output "About to run N skill-forge-judge evaluations — this may take several minutes. Proceed? (y/n)" and stop if denied.
48
49	If a skill's judge run errors: record `ERR` and the error message for that skill; continue to the next skill. Do not abort the audit.
50
51	### Phase 3 — Render Report
52
53	Before rendering, ask: does every layout decision put highest-priority work at the top? If the Needs Work section is empty, lead with a one-sentence "All skills passing" summary instead of an empty header.
54
55	Sort results by grade ascending: F → D → C → B → A. Worst grades appear first — the report is a priority queue, not an alphabetical list.
56
57	Split skills into two groups:
58	- Needs Work: grade below B (< 80%)
59	- Passing: grade B or above (≥ 80%)
60
61	For the `<YYYY-MM-DD>` report header, use the current date already present in the session context — do not shell out for it (`$(date)` and command substitution trigger a permission prompt under this repo's bash rules).
62
63	Output a single markdown report:
64
65	````markdown
66	# Skill Forge Audit
67	_<YYYY-MM-DD> — <N> skills evaluated_
68
69	## Grade Summary
70
71	\| Skill \| Grade \| Score \|
72	\|-------\|-------\|-------\|
73	\| skill-forge-X \| F \| 55/120 (46%) \|
74	\| skill-forge-Y \| C \| 78/120 (65%) \|
75	\| skill-forge-Z \| B \| 98/120 (82%) \|
76
77	## Needs Work
78
79	### skill-forge-X — F (55/120, 46%)
80
81	1. <improvement 1 from judge>
82	2. <improvement 2>
83	3. <improvement 3>
84
85	### skill-forge-Y — C (78/120, 65%)
86
87	1. <improvement 1>
88	2. <improvement 2>
89
90	## Passing (B+)
91
92	\| Skill \| Grade \| Score \|
93	\|-------\|-------\|-------\|
94	\| skill-forge-Z \| B \| 98/120 (82%) \|
95
96	## Skipped
97
98	\| Directory \| Reason \|
99	\|-----------\|--------\|
100	\| some-dir \| No SKILL.md found \|
101	````
102
103	If all skills pass: omit the "Needs Work" section entirely.
104	If no skills pass: omit the "Passing" section.
105	If nothing was skipped: omit the "Skipped" section.
106
107	---
108
109	## NEVER
110
111	- NEVER display full judge reports inline for each skill
112	Instead: Extract only grade, score, and top 1–5 numbered improvements per skill.
113	Why: Unfiltered judge output for 6+ skills floods the context window and buries the aggregate signal the report exists to surface.
114
115	- NEVER apply any fixes during an audit run
116	Instead: Output the report only. Direct the user to `/skill-forge-hitl` or `/skill-forge-update` for remediation.
117	Why: Mixing diagnosis with treatment makes it impossible to know what the baseline was — the audit's value is the before-state snapshot.
118
119	- NEVER sort by skill name or alphabetically
120	Instead: Sort by grade ascending (F first, A last) within each section.
121	Why: Alphabetical sort buries the worst skills; the report's job is to surface where effort is needed most.
122
123	- NEVER read SKILL.md files via `~/.claude/skills/` paths
124	Instead: Use the absolute path to the skills project directory (the repo root where this audit runs).
125	Why: `~/.claude/skills/` is a symlink back to the project; Claude Code'

People who liked this also viewed

6 picks

Type

skill-forge-audit

17stars0forks

View on GitHub

How to install skill-forge-audit?

wrathza/skillforge/skill-forge-audit

$npx -y skills add WrathZA/skillforge --skill skill-forge-audit

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# Skill Forge Audit
2
3	Batch-run skill-forge-judge across every skill in the repo. One consolidated report — grades, top issues, priority order. Read-only: no fixes applied.
4
5	Requires `skill-forge-judge` to be installed.
6
7	No `references/` — this skill is single-file by design; the whole workflow fits in the body, so the absent directory is intentional, not missing progressive disclosure.
8
9	---
10
11	## Workflow
12
13	### Phase 1 — Discover
14
15	Glob for all SKILL.md files, run from the repo root. Skills live one level deep under `skills/`:
16
17	```
18	pattern: skills/*/SKILL.md
19	```
20
21	Collect the directory name of each match — that's the skill name.
22
23	To detect skipped directories: also Glob `skills//` to list every skill directory. Any directory in the `skills//` result that is not in the `skills/*/SKILL.md` result has no SKILL.md — list it in the Skipped section of the report with reason "No SKILL.md found".
24
25	If zero SKILL.md files found: output "No skills found in this repo." and stop.
26
27	### Phase 2 — Evaluate
28
29	For each discovered skill, in order:
30
31	1. Read the SKILL.md file at its absolute path.
32	2. Invoke `/skill-forge-judge` on it.
33	3. From the judge output, extract and record:
34	- Skill name
35	- Grade (A/B/C/D/F)
36	- Score (X/Y, Z%)
37	- Numbered Improvements: up to 5 items (all if fewer than 5 exist)
38
39	Emit a one-line status per skill as each evaluation completes — do not wait until all are done:
40	```
41	✓ skill-forge-create B (98/120, 82%)
42	✓ skill-forge-hitl C (90/120, 75%)
43	```
44
45	If `/skill-forge-judge` is not installed: output "skill-forge-judge is required — install it before running skill-forge-audit." and stop.
46
47	If N > 10: output "About to run N skill-forge-judge evaluations — this may take several minutes. Proceed? (y/n)" and stop if denied.
48
49	If a skill's judge run errors: record `ERR` and the error message for that skill; continue to the next skill. Do not abort the audit.
50
51	### Phase 3 — Render Report
52
53	Before rendering, ask: does every layout decision put highest-priority work at the top? If the Needs Work section is empty, lead with a one-sentence "All skills passing" summary instead of an empty header.
54
55	Sort results by grade ascending: F → D → C → B → A. Worst grades appear first — the report is a priority queue, not an alphabetical list.
56
57	Split skills into two groups:
58	- Needs Work: grade below B (< 80%)
59	- Passing: grade B or above (≥ 80%)
60
61	For the `<YYYY-MM-DD>` report header, use the current date already present in the session context — do not shell out for it (`$(date)` and command substitution trigger a permission prompt under this repo's bash rules).
62
63	Output a single markdown report:
64
65	````markdown
66	# Skill Forge Audit
67	_<YYYY-MM-DD> — <N> skills evaluated_
68
69	## Grade Summary
70
71	\| Skill \| Grade \| Score \|
72	\|-------\|-------\|-------\|
73	\| skill-forge-X \| F \| 55/120 (46%) \|
74	\| skill-forge-Y \| C \| 78/120 (65%) \|
75	\| skill-forge-Z \| B \| 98/120 (82%) \|
76
77	## Needs Work
78
79	### skill-forge-X — F (55/120, 46%)
80
81	1. <improvement 1 from judge>
82	2. <improvement 2>
83	3. <improvement 3>
84
85	### skill-forge-Y — C (78/120, 65%)
86
87	1. <improvement 1>
88	2. <improvement 2>
89
90	## Passing (B+)
91
92	\| Skill \| Grade \| Score \|
93	\|-------\|-------\|-------\|
94	\| skill-forge-Z \| B \| 98/120 (82%) \|
95
96	## Skipped
97
98	\| Directory \| Reason \|
99	\|-----------\|--------\|
100	\| some-dir \| No SKILL.md found \|
101	````
102
103	If all skills pass: omit the "Needs Work" section entirely.
104	If no skills pass: omit the "Passing" section.
105	If nothing was skipped: omit the "Skipped" section.
106
107	---
108
109	## NEVER
110
111	- NEVER display full judge reports inline for each skill
112	Instead: Extract only grade, score, and top 1–5 numbered improvements per skill.
113	Why: Unfiltered judge output for 6+ skills floods the context window and buries the aggregate signal the report exists to surface.
114
115	- NEVER apply any fixes during an audit run
116	Instead: Output the report only. Direct the user to `/skill-forge-hitl` or `/skill-forge-update` for remediation.
117	Why: Mixing diagnosis with treatment makes it impossible to know what the baseline was — the audit's value is the before-state snapshot.
118
119	- NEVER sort by skill name or alphabetically
120	Instead: Sort by grade ascending (F first, A last) within each section.
121	Why: Alphabetical sort buries the worst skills; the report's job is to surface where effort is needed most.
122
123	- NEVER read SKILL.md files via `~/.claude/skills/` paths
124	Instead: Use the absolute path to the skills project directory (the repo root where this audit runs).
125	Why: `~/.claude/skills/` is a symlink back to the project; Claude Code'

People who liked this also viewed

6 picks

Type