skillforge
Audit the whole skill library for agentic readiness and instrument it with telemetry in bulk. USE WHEN auditing skills for autonomy, finding agentic gaps, or adding execution logging across skills. NOT FOR creating a single skill (use CreateSkill) or re-tuning for a new model (use ModelUpgrade).
Source facts
- Repository
- RohiRIK/skills
- Last source activity
- October 16, 2026 at 13:03
- Detected SKILL.md language
- English
- Stars
- 1
- Forks
- 1
Install options
The review-first prompt is selected by default. You can switch to a direct command or download a local copy.
Review the source files
Read SKILL.md and any companion files shown by SkillsMP before deciding whether to install.
File Explorer
4 filesShowing SKILL.md
SKILL.md
Source instructions · Read-only preview- name
- SkillForge
- description
- Audit the whole skill library for agentic readiness and instrument it with telemetry in bulk. USE WHEN auditing skills for autonomy, finding agentic gaps, or adding execution logging across skills. NOT FOR creating a single skill (use CreateSkill) or re-tuning for a new model (use ModelUpgrade).
- category
- meta
- effort
- medium
- domain
- meta
- tuned-for
- claude-5.5
- tuned-on
- 2026-10-14T00:00:00.000Z
# SkillForge
Fleet-level meta-skill for the skill library. Where `CreateSkill` operates on **one skill at a time** (scaffold, validate, test, improve), SkillForge operates on the **whole library at once** — scoring every skill for agentic capability and instrumenting them in bulk.
It does not create skills. Authoring a new skill is `CreateSkill`'s job (and the upgraded `CreateSkill` already makes new skills loop-ready). SkillForge answers a different question: across everything I already have, what is missing and what needs instrumenting?
## Workflow Routing
| Workflow | Trigger | File |
|----------|---------|------|
| **AuditAgentic** | "audit skills for autonomy", "agentic gaps", "readiness report" | `Workflows/AuditAgentic.md` |
| **AddTelemetry** | "add execution logging", "instrument skills" | `Workflows/AddTelemetry.md` |
## Quick Reference
- Agentic-readiness rubric (5 axes): Iteration, Verification, Reflection, State, Telemetry
- Scope is the whole library — fleet audit, not single-skill authoring (that's `CreateSkill`)
- Pairs with `CreateSkill` (builds one skill) and `Prompting` (wording); SkillForge measures + instruments the set
- Shared state lives in `.agent-state.md`; telemetry in `~/.claude/state/execution.jsonl`
## Gotchas
- Tier-A reference skills are "reference, exempt" — never score them as single-shot failures.
- Telemetry presence is per action workflow, not per skill.
- This audits/instruments the fleet; authoring one skill is `CreateSkill`'s job.
## Examples
**Example 1: Score the whole library**
```
User: "audit my skills for autonomy"
→ Invokes AuditAgentic workflow
→ Scores every skill on the 5-axis readiness rubric
→ Ranked report: which skills are reference-exempt, single-shot, partial, or loop-ready
```
**Example 2: Instrument existing skills in bulk**
```
User: "add execution logging to all my skills"
→ Invokes AddTelemetry workflow
→ Appends the standard JSONL execution-log line to each action workflow that lacks it
```
**Example 3: Find what to fix next**
```
User: "where are the agentic gaps in my library?"
→ AuditAgentic → top fixes ranked by how many skills each change unblocks
```
View on GitHub