designing-workflow — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited designing-workflow (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
This skill defines the canonical six-phase workflow menu for rageatc-core-oss and provides guidance on phase selection based on task characteristics. All workflows subset from this menu.
Scope: This skill covers workflow structure (what phases exist and when to use them). For workflow execution (agent coordination, logging, validation), see orchestrating-work skill.
Use this skill when:
Do NOT use for:
orchestrating-work)orchestrating-work)orchestrating-work)orchestrating-work)All rageatc-core-oss workflows subset from these six phases. Custom workflows select relevant phases based on task characteristics.
Purpose: Understand requirements, explore context, identify execution approach.
When to include: Always. This is the foundational phase.
Key activities:
ideating skill to generate, capture, and explore ideas when the user doesn't yet have a clear task (optional, explicitly invoked)understanding-the-ask skill to explore user intent through structured dialoguesolutioning skill to explore solution options when approach is unclear (optional, explicitly invoked)Outputs: Clear understanding of requirements, directional ideas on execution. Decision record if solutioning was used. Captured ideas if ideation was used.
Notes: This is where orchestrator builds shared understanding with user before planning execution. The full Preparation arc is: Ideate → Understand → Solution → Brief (each optional). The ideating skill helps when the user needs to generate or explore ideas before committing to a direction. The understanding-the-ask skill provides frameworks for structured questioning and readiness criteria. The solutioning skill helps when the problem is clear but the solution direction needs exploration.
Purpose: Create thorough execution plan for orchestrator and user.
When to include: Always. This is where most time should be spent with user.
Key activities:
Outputs: Plan document (work/<task-id>/plan.md) with phase sequence, decision points, dependencies, file paths.
Notes: Plan is for orchestrator and user, NOT for agents (they receive instructions at invocation). If planning is done right, everything else flows smoothly.
Purpose: Background investigation, fact-finding, evidence gathering.
When to include:
When to skip:
Sub-steps:
Outputs: Research document, source index (thorough mode), fact-check report.
Notes: Research is the primary fact-checking checkpoint. Verify claims here to prevent error propagation to artefacts.
Purpose: Create primary deliverable through iterative refinement.
When to include: Always. This is the core value-creation phase.
Sub-steps:
Outputs: Artefact versions (v1, v2, v3...), review versions (matching artefact versions), approved final version.
Notes: Iteration occurs within Production through Producer-Critic cycles (sub-phase iteration). This is low-cost quality refinement, not phase-level rework.
Purpose: Extract system improvements from iteration patterns, identify skill gaps, propose workflow enhancements.
When to include:
When to skip:
Key activities:
Outputs: Learning proposals document.
Notes: Learning comes BEFORE Finalisation. Proposals are reviewed and approved by user before updating skills or workflows.
Purpose: Deploy results, update project state, make commits.
When to include:
When to skip:
Key activities:
Outputs: Deployed artefact, updated project files.
Notes: Always the last step. Human approval gates typically belong here (at irreversible actions like deployment or commits).
Three common patterns emerge from phase selection.
Phases: Preparation → Planning → Production → Finalisation
Use when: Requirements clear, domain familiar, low iteration expected.
Example: Update documentation with known changes.
Phases: Preparation → Planning → Research → Production → Finalisation
Use when: Unfamiliar domain requires investigation, evidence needed for decisions, factual claims need verification.
Example: Create skill for new domain requiring background research.
Phases: Preparation → Planning → Research → Production → Learning → Finalisation
Use when: High-stakes work, building new capabilities, complex work with learning opportunities.
Example: Create foundational skill with novel patterns requiring research and learning extraction.
Use this framework to select phases for a task.
Step 1: Start with mandatory phases
Step 2: Evaluate Research phase
Step 3: Evaluate Learning phase
Result: Phase sequence (e.g., Preparation → Planning → Research → Production → Learning → Finalisation)
Default when uncertain: Use Research-Driven pattern (include Research, skip Learning). Research prevents costly Production rework; Learning can be added post-hoc if valuable patterns emerge.
Seven principles govern effective workflow design:
Workflows must specify context handling:
What persists between phases:
What resets between phases:
Handoff protocols: Workflow design specifies what context transfers at phase boundaries. See orchestrating-work for how to log and validate handoffs during execution.
Phase boundaries requiring explicit handoffs:
This skill defines workflow structure; orchestrating-work defines execution.
designing-workflow provides:
orchestrating-work provides:
Usage pattern:
designing-workflow to plan phase sequenceorchestrating-work to coordinate agent execution within phasesA workflow design is complete when:
Workflow design is effective when:
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.