assess-findings
Assess Findings
How to assess findings — whether from an AI reviewer, an auditor, a challenger, or any other source of AI-generated feedback. This protocol applies to every context where an agent must decide which findings are correct, which are noise, and what to do about each.
Stance
Findings are hypotheses about the work. Some will be correct and important. Some will be wrong, inapplicable, or based on incomplete understanding. Your job is to determine which is which, with evidence.
Do not default to acceptance. The path of least resistance (apply every suggestion) produces worse code than thoughtful evaluation. Equally, do not default to rejection — that wastes valid insights.
Assess each finding by merit. The question is not "is this a big issue?" or "is this convenient to fix?" The question is: is the finding correct? A correct improvement is worth implementing regardless of size. An incorrect suggestion should be declined regardless of who raised it.
Phase 1: Investigate
For each finding, build enough context to make a high-confidence judgment. Do not assess from the finding description alone.
More from inkeep/team-skills
qa
Manual QA testing — verify features end-to-end as a user would, by all means necessary. Exhausts every local tool: browser (Playwright), Docker, ad-hoc scripts, REPL, dev servers. Mock-aware — mocked test coverage does not count. Proves real userOutcome at highest achievable fidelity. Blocked scenarios flow to /pr as pending human verification. Standalone or composable with /ship. Triggers: qa, qa test, manual test, test the feature, verify it works, exploratory testing, smoke test, end-to-end verification.
61spec
Drive an evidence-driven, iterative product+engineering spec process that produces a full PRD + technical spec (often as SPEC.md). Use when scoping a feature or product surface area end-to-end; defining requirements; researching external/internal prior art; mapping current system behavior; comparing design options; making 1-way-door decisions; negotiating scope; and maintaining a live Decision Log + Open Questions backlog. Triggers: spec, PRD, proposal, technical spec, RFC, scope this, design doc, end-to-end requirements, scope plan, tradeoffs, open questions.
54cold-email
Generate cold emails for B2B personas. Use when asked to write cold outreach, sales emails, or prospect messaging. Supports 19 persona archetypes (Founder-CEO, CTO, VP Engineering, CIO, CPO, Product Directors, VP CX, Head of Support, Support Ops, DevRel, Head of Docs, Technical Writer, Head of Community, VP Growth, Head of AI, etc.). Can generate first-touch and follow-up emails. When a LinkedIn profile URL is provided, uses Crustdata MCP to enrich prospect data (name, title, company, career history, recent posts) for deep personalization.
54docs
Write or update documentation for engineering changes — both product-facing (user docs, API reference, guides) and internal (architecture docs, runbooks, inline code docs). Builds a world model of what changed and traces transitive documentation consequences across all affected surfaces. Discovers and uses repo-specific documentation skills, style guides, and conventions. Standalone or composable with /ship. Triggers: docs, documentation, write docs, update docs, document the changes, product docs, internal docs, changelog, migration guide.
52implement
Convert SPEC.md to spec.json, craft the implementation prompt, and execute the iteration loop via subprocess. Use when converting specs to spec.json, preparing implementation artifacts, running the iteration loop, or implementing features autonomously. Triggers: implement, spec.json, convert spec, implementation prompt, execute implementation, run implementation.
52ship
Orchestrate any code change from requirements to review-ready branch — scope-calibrated from small fixes to full features. Composes /spec, /implement, and /research with depth that scales to the task: lightweight spec and direct implementation for bug fixes and config changes, full rigor for features. Produces tested, locally reviewed, documented code on a feature branch. The developer pushes the branch and creates the PR. Use for ALL implementation work regardless of perceived scope — the workflow adapts depth, never skips phases. Triggers: ship, ship it, feature development, implement end to end, spec to PR, implement this, fix this, let's implement, let's go with that, build this, make the change, full stack implementation, autonomous development.
52