Traction scoreGitHub stars can be faked, so popularity alone can be misleading. Traction Score looks for broader signs of recent attention, adoption, and active maintenance.
This collection provides tools for AI agent builders to systematize the transition from raw task execution to structured capabilities. By analyzing browser flows and artifact pipelines, it allows developers to capture successful run histories as new skill packages.
The repository focuses on the iterative refinement of agent behavior, using validation failures and real-world evidence to apply structural upgrades to existing skills.
WHO IT'S FOR
AI agent / automation builders
creating reusable skill packages from run history
AI workflow engineers
optimizing agent skills based on failure traces
Technical recruiters / hiring managers
automating interview candidate assessment and reporting
Academic writers / researchers
humanizing academic content for better readability
Compatible AgentsThe repository documents support for these agents. The skills may also work with other agents that can load SKILL.md files, but they may need some setup or small changes.
Installs: Complete COMPASS Skills bundle containing all nine documented skill folders and their supporting subdirectories.
Before you start
Python 3.10 or later must already be installed; the complete bundle includes assess-interview-candidate, which requires it.
Manually copy all nine folders under the repository's skills/ directory into the agent's local skills directory. Keep every references/, scripts/, assets/, evals/, and agents/ subdirectory intact.
Replace “Task: ...” with the task that needs clarification.
Replace “Material: ...” with the relevant text, file paths, or context, or state that no material is provided.
Check the repository for this skill’s setup.
Use $task-clarifier to align the task below.
Task: ...
Material: ...
Constraints: ask user-owned decisions first; infer discoverable facts from files, context, or reliable sources. Ask only questions that change scope, method, evidence, format, safety, or acceptance criteria.
Output: ask 1-3 key questions with recommended answers first; once the core need is clear, restate your understanding in 2-5 lines and ask me to confirm.
The current AI conversation with enough project context to identify goals and progress.
Explicit approval is required before applying the proposed changes.
Check the repository for this skill’s setup.
Use $task-forest to analyze the current AI conversation and maintain the task forest for this workspace.
Goal: create a task-forest proposal from long-running goals, tasks, progress, deviations, risks, decisions, and follow-ups in this AI conversation.
Requirements:
1. Read the current task-forest list and todo first; initialize task-forest if missing.
2. Identify which long-term goal this AI conversation served. If no relation is clear, ask me or create a question/risk node.
3. Save a proposal and show me the planned changes before applying.
4. After approval, apply, validate, export, and report the HTML path.
Optionally, the desired focus for the next session and any explicit files or local transcript paths that may be used.
Check the repository for this skill’s setup.
Use $session-handoff-prompt to create a concise, paste-ready continuation prompt for a fresh agent session. Use the current visible conversation and workspace evidence needed to preserve the objective, hard requirements, verified facts, decisions, completed work, pending work, key files, risks, and next actions. Keep the prompt local to this machine and label facts as verified, inferred, or unverified. Do not execute any remaining work.
Claude
Codex
OpenClaw
OpenCode
Hermes