agentleFS
lalulalilalulali/vespyr/.agents/skills/discovery-report/SKILL.md

discovery-report

Skill11 starsChanged 11 days agoMITView on GitHub

Compile design thinking and research outputs into a unified report with dynamic usability testing scoring. Use when the user says "compile everything we learned", "pull the research into one report", "write the discovery report", or invokes /discovery-report before presenting findings to stakeholders.

1 other project uses a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

4 rules. About 0 are common, 4 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    After completing multiple design thinking exercises

    line 19
  2. Distinctive:

    Before presenting findings to stakeholders

    line 20
  3. Distinctive:

    Write the appropriate output file, following the canonical template:

    line 51
  4. Distinctive:

    every present gated input to actually PASS its gate, re-verified at compile time — an artifact can change after it was written, so a green …

    line 82

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Discovery Report

What this skill does

Compiles research and design thinking outputs into a single unified report. Dynamically includes or excludes the Usability Testing (UT) Score section based on available data. Outputs to artifacts/output/02-research/.

Persona delegation

This skill delegates to @product-manager. The pm compiles and synthesizes findings. The skill provides the report structure; @product-manager provides the narrative synthesis and prioritization.

When to use

  • After completing multiple design thinking exercises
  • Before presenting findings to stakeholders
  • As a handoff from research to design
  • "Compile everything we've learned into one report"

Workflow

Step 1: Inventory available artifacts

Check artifacts/output/02-research/ for:

  • problem-space-brief.md
  • root-cause-analysis.md
  • research-plan.md
  • empathy-map.md
  • journey-map.md
  • jtbd-hmw.md
  • Usability testing data (UT scores)

Step 2: Compile report

Synthesize findings into a unified narrative. Focus on connections: how does the root cause connect to the journey friction? How do JTBDs map to empathy pains?

If usability testing data exists:

  • Output usability-report.md with UT scores, task completion rates, error rates, and satisfaction metrics
  • Include the SUS (System Usability Scale) score if calculated

If no usability testing data:

  • Output user-research-report.md without the UT sections — but keep Research Methods Used and Open Questions (see Step 3)
  • Note: "Usability testing not yet conducted. Recommend running usability tests before finalizing design specs."

Step 3: Output

Write the appropriate output file, following the canonical template:

.agents/templates/discovery/discovery-report.md

That template is the single source of truth for the report shape. Do not inline a copy here — a copy drifts, and a drifted copy is how a report ends up shipping without its methods or its open questions.

The template handles both branches; you choose the output filename and drop the sections that have no data:

SituationOutput fileTemplate sections to include
UT data existsartifacts/output/02-research/usability-report.mdAll of them, including ## Usability Testing Results (if available) and ## SUS Score (if available). Title: "Usability Report"
No UT dataartifacts/output/02-research/user-research-report.mdOmit both UT sections. Keep ## Research Methods Used and ## Open Questions — they are what makes the report reviewable. Title: "User Research Report"

Never drop ## Research Methods Used or ## Open Questions because there is no UT data. They are the provenance and the unresolved-questions record; a report without them cannot be challenged, only believed. When no UT data exists, add this line under Recommendations rather than removing a section:

Usability testing not yet conducted. Recommend running usability tests before finalizing design specs.

Done when — checked mechanically

This skill compiles five discovery artifacts into one report for stakeholders. Four were gated (discovery_canvas_check.js) and the fifth (research-plan.md) is gated too. The compiler had zero gate references — which is the worst place for a gap: it consumed verified inputs and emitted an unverified output. A reader of the consolidated report had no assurance that ANY underlying artifact conformed, because the report laundered their status.

node .agents/scripts/discovery_report_check.js check \
  --report artifacts/output/02-research/user-research-report.md --root . --json

It requires:

  • every present gated input to actually PASS its gate, re-verified at compile time — an artifact can change after it was written, so a green result from an earlier run is not evidence about the file on disk now
  • a ## Sources section naming each input with a machine-readable status: PASS, FAIL, absent, or not-gated
  • every absent input to be disclosed. An undisclosed absence means a reader cannot tell missing research from research that found nothing
  • the usability caveat when no usability data was compiled. The skill instructs the author to note it in prose; nothing checked it

Measured, the laundering it stops: a report whose Sources row claimed PASS for an empathy-map.md that was verbatim its own unfilled template — 8 placeholder failures — was reported to stakeholders as a verified finding.

What it does not check: whether the synthesis is any good, whether the connections drawn between findings are real, or whether the report is fit to show a stakeholder. A gate cannot verify a narrative. What it can do is stop the narrative implying a verification that never happened.

Known gap — declared inputs with no producer

problem-space-brief.md and usability testing data are declared inputs that no skill in this repo produces. They are disclosed rather than required, deliberately: a hard requirement with no producer makes the skill permanently unsatisfiable, which is the qa-signoff.md deadlock. Usability testing in particular is a real capability gap, not a wiring bug — nothing here runs a usability study.

Note on output naming: this skill emits usability-report.md or user-research-report.md, depending on whether usability data exists. They are mutually exclusive outputs of this skill, not inputs — and a consumer must read the skill to know which file to open.

Exit 0 conformant · 1 failures · 2 usage/IO error, fail-closed.

State machine integration

At start: Run node .agents/scripts/orchestrator_state.js status At end: Run node .agents/scripts/orchestrator_state.js complete --agent product-manager --artifact 02-research/{report-type}.md

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.