⏳ This skill is pending AI review.
Scores will appear once the review pipeline completes.
ocas-forge
Skill architect and builder. Designs, builds, and validates complete Agent Skill packages through a mandatory eight-phase pipeline. Default output is the finished installable package. Not for skill evaluation (use skilllab) or variant proposals (use ocas-mentor).
Choose how to use this skill
You do not need every option. Choose the path your AI client supports. The stable page stays the same; versioned files are immutable.
1. Native installer
This listing has no registered native installer command. Use the complete package or source fallback below, depending on what your client supports.
Do not guess an installer command or replace an existing version without reviewing the diff.
2. Complete package recommended
Download the ZIP when available. It includes SKILL.md plus the references, security notes and version metadata.
No complete ProSkills package is published for this listing yet.3. Prompt-only
Copy the prompt above when the agent can read the stable page or when you want to adopt the workflow without installing a skill.
Need only the instruction file?
Download SKILL.md only if your client requires a single file. The complete ZIP is safer for a full installation because it preserves the references and release context.
No path installs or executes anything by itself. Your agent still needs access to the project files. Before updating, compare the installed version and review the diff.
// RATINGS
Not yet listed on ClawHub or SkillsMP
// README
Forge
Skill design and construction for Hermes Agent. Runs a mandatory eight-phase pipeline: existence gate, research, classify, scope, architecture, plan, build, validate. Default output is the finished installable package.
Skill name: ocas-forge
Version: 3.7.1
Type: system
Layer: Architecture
Author:
📖 Overview
Builds and repairs agent skill packages. Default output is the finished package.
🔧 Commands
forge.build— build a complete skill packageforge.critique— review a package and identify defectsforge.repair— fix broken files in an existing packageforge.classify— classify a proposed skill by size classforge.validate— run structural validationforge.scaffold— generate a minimal package skeletonforge.consolidate— merge orphan or duplicate skill into its natural parentforge.sync— sync local skill changes to the canonical repository via PRforge.verify-update— check whether a skill is at the latest versionforge.audit— audit skills for compliance, apply fixes, sync to GitHubforge.status— current build state if multi-step build in progressforge.journal— write journal for the current runforge.update— pull latest from GitHub source; preserves journals and data
📊 Outputs
Packages written to disk. Journals. Validation reports. PRs for sync operations.
📄 Files
| File | Purpose |
|---|---|
SKILL.md | Skill definition with agentskills.io frontmatter |
CHANGELOG.md | Release history |
references/ | Pipeline specs, authoring rules, ontology, schemas, workflow plans |
scripts/ | Build and validation scripts |
templates/ | Package scaffolding and file templates |
evals/ | Evaluation suites and metrics |
deferred/ | Out-of-scope material pending proper homes |
Changelog
- v3.7.1 (2026-07-13) — Audit repair: restored missing manifest fields, fixed frontmatter validation errors, standardized templates/skill.json against schema.
- v2.7.1 (2026-04-26) — Version alignment: SKILL.md frontmatter, CHANGELOG.md, and GitHub release tag in sync. No functional change.
- v2.7.0 (2026-04-18) — Refactor: folded builder-domain helper skills into Forge's existing structure. SKILL.md shrunk to 343 lines. Added
forge.consolidate,forge.sync,forge.verify-update. Moved runtime-harness guidance todeferred/running-ocas-skills.md. Moved OAuth/MCP setup content todeferred/mcp-oauth-setup.md. - v2.6.11 (2026-04-16) — Authoring rules v2.7.2 -> v2.8.0. Ontology v2.0.1 -> v2.1.1. Active ecosystem updated.
- v2.6.8 (2026-04-12) — Fixed missing
{agent_root}/skills/filesystem write path. - v2.6.0 (2026-04-08) — Adopted agentskills.io open standard. Replaced skill.json hardcoding with {agent_root}/commons/ convention.
📚 Documentation
Read SKILL.md for operational details, schemas, and validation rules.
Read references/ for detailed specifications and examples.
📄 License
MIT License — see LICENSE for details.
// HOW IT'S BUILT
// FILES
2 files shipped with this skill. Inline text is the catalog copy. The upstream link is the source repository.
SKILL.md7.9 KBupstream ↗
--- warning: 'FALSE TRIGGER RISK: rate CORRECTED 2026-09-26 from 100% (2/3) to NOT-MEASURED. All 3 interactive loads were user_prompted=true, i.e. explicitly requested, so an interactive false-trigger rate is not meaningful for this skill. The prior 100% also carried a miner bug that attributed the next 10 messages'' tools to every skill loaded in the same burst. Do not treat this skill as excluded — it is safe to load on request.' name: ocas-forge description: 'Skill architect and builder. Designs, builds, and validates complete Agent Skill packages through a mandatory eight-phase pipeline. Default output is the finished installable package. Not for skill evaluation (use skilllab) or variant proposals (use ocas-mentor).' license: MIT source: https://github.com/indigokarasu/forge includes: - references/** - scripts/** metadata: author: Indigo Karasu (indigokarasu) version: "3.8.0" hermes: category: software-development tags: - skill-builder - architecture - design - OCAS-core tags: - skill-builder - architecture - design - OCAS-core triggers: - build a skill - create skill - skill architecture - design skill package - skill builder --- # Occas-Forge Skill architect and builder. Given a capability idea or broken package, runs a mandatory eight-phase pipeline before writing files. Default output is the finished installable package — never returns design briefs or plans in place of real artifacts. Every build starts with absorption check, research, and classification before a single file is created. **Support files:** `references/support-file-map.md` indexes the bundled files not covered inline in this skill — check it before working from assumptions about what is (not) available. ## Why This Skill Exists Forge is the only authorized skill builder in OCAS. Without it, agents would create skills with inconsistent structure, missing frontmatter, no tests, and no cross-skill coordination. The rigid pipeline ensures every skill meets OCAS standards and doesn't duplicate existing work. ## When to Use - Building new OCAS skills from scratch - Skill architecture and design review - Bulk skill library updates and synchronization - Skill consolidation and deprecation - Repair broken or defective skill packages - Validate a skill package against OCAS standards ## When NOT to Use - **Skill evaluation/scoring** — use skilllab - **Variant proposals/experiments** — use Mentor or Fellow - **System health monitoring** — use Custodian - **One-off task execution** — use the appropriate existing skill - **Web research** — use Sift - **Authentication and MCP wiring** — use ocas-auth - **Frontmatter reference/template lookup** — use `write-a-skill` for quick field requirements - **Building non-skill artifacts** ## Pipeline - [ ] **Existence gate** — check if skill already exists, absorption test - [ ] **Research** — search GitHub + skill library for patterns; review ≥10 sources - [ ] **Classify** — determine shortcut/workflow/system type - [ ] **Scope** — define boundaries and interfaces - [ ] **Architecture** — design package structure - [ ] **Plan** — create implementation plan - [ ] **Build** — implement the skill - [ ] **Validate** — run critique and verify 50/50 **Absorption first:** If an existing skill already owns the domain, add content as a `references/` doc or `scripts/` file — do NOT create a new skill. See `references/enforcement_durability.md`. **Research is mandatory for ALL operations** — not just new builds. When improving an existing skill, you MUST research external sources for new patterns. **Pre-Build Quality Linters** (embedded in the Validate step, per [[`spec-ocas-skill-improvements.md` ⚠️ Pending spec] ⚠️ Pending spec — not yet authored] §5.1): 1. **Frontmatter verification** — validate required metadata fields (`name`, `description`, `version`, `author`). 2. **Incident-log shape detection** — flag skill descriptions / reference bodies overly dense in ephemeral issue numbers or quoted chat transcripts instead of generalizable rules. 3. **Reference sprawl check** — fail the build if `references/` exceeds 60 files. 4. **Config vs env separation** — reject scripts that read behavioral configuration from `os.environ` instead of `skills.config.<key>` in `config.yaml` (reserving env vars strictly for credentials/secrets). Run all four before a new package is marked valid; a failing linter blocks promotion to production. ## Commands - `forge.build` — design, scope, build, validate a complete skill package - `forge.critique` — review package and identify defects - `forge.repair` — fix broken files in existing package - `forge.validate` — run validation checks on a package - `forge.audit` — audit one or more skills for OCAS compliance, apply fixes, sync to GitHub - `forge.consolidate` — merge orphan/duplicate skill into natural parent - `forge.sync` — sync local changes to canonical repository via PR - `forge.update` — pull latest from GitHub source - `forge.status` — current build state (multi-step) - `forge.journal` — write journal for current run **Run completion:** After every command, check for unprocessed VariantProposal/VariantDecision files in `intake/`, process new files, persist build logs to `decisions.jsonl`, write journal. Never report success until validation has actually executed — `write_file`/`skill_manage(action='create')` completing is NOT validation. ## Triage Classification (dispatch) | Intent | Priority | Action | When | |--------|----------|--------|------| | new_journals | 50 | dispatch_scan | Files need eval-store bridge | | explicit_run | 80 | run_full_pipeline | Explicit-run override detected | | re_detection | 30 | closure_noop | All journals already in both eval stores | ## Skill Type Classification - **Shortcut** — narrow tool wrapper. 20-120 line SKILL.md. - **Workflow** — multi-step process. 80-250 line SKILL.md. - **System** — durable behavior system. 150-300 line SKILL.md + references. ## Error Handling | Error | Handling | |-------|----------| | `closure_closeout_check.py` gate [1] reads `dispatch=False` although your audit found the journal bridged in both stores | `in_store()` (line ~54) is a raw substring grep for the **full relpath** in BOTH `ocas-praxis` and `ocas-dispatch` `journals_evaluated.jsonl`. An entry carrying only `filename: <bare>.json` matches Step 2 / human greps but NOT gate [1]. Emit both keys in the dispatch entry: `filename` (bare) and `journal_id` (`ocas-<skill>/<date>/<file>.json`). Do not trust a `GAPS=0` from a basename-matched audit — audit with the reader's own semantics. | | `closure_closeout_check.py` reports `[2] monitor ROOT ... : False (MISSING ...)` while `advance_gate_state.py` wrote a real file | The checker's `MON_STATE_ROOT` is derived via `HERMES_HOME.replace("/profiles/indigo", "/commons/data/...")`. When `HERMES_HOME=/root/.hermes` (no `/profiles/indigo` segment) the replace is a no-op and the path resolves to a nonexistent file, so the gate reads MISSING even though the real monitor copy is advanced. Do NOT re-advance and do NOT report STALE. Re-run with `HERMES_HOME=/root/.hermes/profiles/indigo` so the replace matches and the check reads the file that was actually written. Requires `=== gates ALL CLOSED ===`. | | `closure_closeout_check.py` crashes (FileNotFoundError – `<hermes-home>` placeholder) | Skip script. Manually verify gates: (1) grep bare relpath in both eval stores, (2) recompute max journal mtime programmatically + ≥2s pad, advance both monitor copies + praxis `ingest_state.json`, (3) re-assert `verified_second_wave` on all dispatch-owned files. | | `run_mixed_wave_closure.py` crashes (NameError: 'os') | Handle closure manually. Same manual procedure as above. | | `bridge_eval_inline.py` crashes (FileNotFoundError) | Manual JSONL append to both eval stores. | | `forge_count_unprocessed.py` returns 0 but dispatcher says genuine | Cross-check against prior `forge-scan-*.json` journal's `unprocessed_proposals` fieldREADME.mdupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
Plain-text copy, no login: /skill-bundles/indigokarasu-forge.md