pesteph
UserA skills-based workflow template for structured AI-assisted development with GitHub Copilot CLI und Claude Code Support for other harnesses will be added based on user demand
Categories
Indexed Skills (19)
install-workflow
Install the agentic-workflow methodology into a target project. Detects existing installs (also old layouts), supports clean installs and upgrades, asks the user about locally modified Skill files. Run this from inside the agentic-workflow repo.
analyze
Analyzes a GitHub issue, opens up the problem space, and generates a /research prompt. Use this Skill when an issue or task needs to be analyzed.
axiom
Defines the universe of the project through Socratic dialogue. Writes the result into docs/. Use this Skill once at the start of a project or when the stack/constraints fundamentally change.
brief
Generates design.md (the green figure) as the handoff artifact for Phase 4 (Build). Always use this Skill at the end of Phase 3 (/design).
conceptualize
Shows options for action and evaluates them. Use this Skill when multiple solution approaches are possible and an informed choice needs to be made — as a decision checkpoint before detailed analysis.
deep-conceptualize
Multi-Agent concept exploration for complex architecture decisions. Alternative to /conceptualize — use this Skill when multiple options should be analyzed deeply in parallel.
design
Detailed analysis of the approved options. Formulate ADRs, create pseudocode as a collaborative artifact. Use this Skill when architecture and technical details should be worked out — after /conceptualize or directly after /analyze for straightforward tasks.
diaboli
Devil's Advocate — attacks design decisions creatively AND verifies the design's completeness systematically. Use this Skill optionally after /design when architecture decisions should be challenged and the plan should be checked before implementation.
discuss
Reveal gray areas, make assumptions explicit, and capture user preferences before the concept. Use this Skill when unclear points or implicit assumptions need to be clarified before an architecture decision.
doc-review
Checks whether the documentation matches the code and the specs. Use this Skill with a PR number to ensure documentation consistency.
downstream
Pulls the latest version of the Skills from the original Skill repo and merges changes into the project. Use this Skill to bring Skills up to date.
qa
Meta-Skill that dispatches all 5 QA reviews in parallel as a Fleet. Use this Skill after implementation to perform quality assurance efficiently.
research
Enriches context with researched information from the codebase, the web, and public repositories. Produces a cited research report. Use this Skill to close knowledge gaps before conceptualizing or designing — e.g. from a research prompt generated by /analyze.
retro
Conducts a workflow retrospective. Identifies what went well, what can be improved, and suggests concrete improvements. Use this Skill at the end of a workflow run.
security-review
Conducts a security review of a pull request. Mandatory STRIDE coverage. Analyzes vulnerabilities, injection risks, auth issues, dependencies, and other weaknesses. Use this Skill with a PR number or file paths.
simplify
Simplifies code in a pull request. Reduces complexity, improves readability, and removes unnecessary abstractions. Use this Skill with a PR number or file paths.
test-review
Calculates test coverage mathematically, performs logic tracing, and checks test conventions. Use this Skill to evaluate the test quality of a change.
upstream
Transfers Skill improvements from /retro to the original Skill repo as a PR. Use this Skill after /retro to share improvements with all Skill consumers.
next
Detects the current workflow state from plan.md and executes the next Skill documented there. Use this Skill to continue the workflow.
Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.