skills

Synced skill library for reusable task recipes.

perceptis-manual-import-connector

Source: droxey/nebula-recipes/tasks/perceptis-manual-import-connector

Sample prompt:

Use the perceptis-manual-import-connector skill to validate a Perceptis export, normalize the rows, and produce preview, duplicates, exceptions, and summary outputs without modifying the source data.

perceptyx-read-only-export-validation-and-scaffold

Source: droxey/nebula-recipes/tasks/perceptyx-read-only-export-validation-and-scaffold

Sample prompt:

Use the perceptyx-read-only-export-validation-and-scaffold skill to review a Perceptyx export workflow, keep the process read-only, and scaffold the connector files needed for validation and normalization.

intake-implementation-workflow

Source: droxey/nebula-recipes/tasks/intake-implementation-workflow

Sample prompt:

Use the intake-implementation-workflow skill to turn an intake request into a planning package with a spec, implementation plan, build plan, and implementation checklist.

musexmachine-mcp-codex-env

Source: musexmachine/mcp

Sample prompt:

Use the musexmachine-mcp-codex-env skill to create and validate a Codex-ready environment for musexmachine/mcp.

code

Source: hacker-news-comment (workflow pattern)

Sample prompt:

Use the code skill to execute a refactor with plan iteration, spec-first thinking, tests-first implementation, manual verification, review, mutation testing, and docs updates.

reasoning-router

Source: reasoning-router/

Sample prompt:

Use the reasoning-router skill before beginning the task to select the lowest sufficient available reasoning mode.

product-reverse-engineering

Source: product-reverse-engineering/

Sample prompt:

Use $product-reverse-engineering to route this product analysis or rebuilding request to the correct specialist with the required safety gates and reviewed handoffs.

skill-maturity

Every first-party skill declares a maturity level in its SKILL.md frontmatter and carries the required Purpose, Inputs, Outputs, Example, Success criteria, and Maturity sections. Levels: 0 Intent, 1 Determinism (script/structured asset), 2 Stability (unit tests), 3 Safety and Scale (agents/ interface or safety section). See docs/skill-maturity-standard.md.

Validate and harden the library:

python3 tools/validate_skills.py          # strict check of a skill directory's frontmatter + sections
python3 tools/harden_skills.py            # batch-add maturity + required sections to live first-party skills
python3 tools/finalize_skills.py          # harden the essence meta-skills and sweep-validate all skills
python3 -m unittest tools.test_validate_skills

Use this repository with any agent harness

Kickoff prompt

Use this repository as a library of reusable skills. Read the relevant
SKILL.md before starting, follow its instructions, and keep the task scoped
to the user's request. If no skill clearly applies, proceed without one.

Usage guide

1. Make this repository available to your harness by cloning it or adding it as a skills directory.
2. For each task, select the closest skill below and provide its SKILL.md to the agent as context.
3. Ask the agent to follow the skill, inspect the repository, and report validation results.

Skills

• ah-devops-engineer
• ai-writing-detection
• chrome-mcp-web-fallback
• code
• code-review
• codespace-operator-v2
• dani-roxberrys-teaching-voice
• essence
• essence-commit
• essence-doc
• essence-pr
• fleet-audit-patterns
• go-defensive
• humanize
• improve
• intake-implementation-workflow
• ios-rendering-protocol
• linkedin-lead-gen-outreach
• markdown-mode-router
• meeting-to-action
• mobile-forensics
• model-cost-estimator
• nextjs-nebula-miniapp
• owasp-top-10
• product-clone-research
• product-reverse-engineering
• prompt-technique-router
• qmd
• repo-settings-bootstrap
• resume-ats-pdf-optimizer
• senior-devops
• skill-router-v2
• tdd-workflows-tdd-cycle
• teaching-lesson-plan
• web-scraper