discovery-synthesize-research

Turn raw research into a synthesis anchored on a local Discovery/Spike work item under docs/discovery/. Use when research has been gathered and needs to become navigable findings: phrasings like "synt

SKILL.md

GitHub

Discovery research synthesis

Owned by the Analyst persona (the synthesize-research step). Invoked as /discovery-synthesize-research.

Turn raw research (interviews, surveys, usability sessions, a study held in a research tool, or a repo doc) into a concise synthesis anchored on a local Discovery/Spike work item, so the findings are navigable from the work-item spine and feed elaboration. This is the entry to the discovery phase.

The anchor model (read before writing)

Research runs in an upstream tool or doc. The local Spike work item under docs/discovery/spikes/ is the authoritative anchor: it holds the link to the upstream study and a short synthesis summary (key findings, implications, open questions). The graph stores the edge (the link) and the summary, not the research content, so a developer reaches the findings from the work item without needing upstream tool access.

Workflow

  1. Gather the research and its source link. Collect the raw inputs and the canonical URL or repo path of the study. If no stable link exists, ask for one - the edge is the whole point.

  2. Synthesise. Distil the research into three parts, kept short and specific:

    • Key findings - what the research actually showed, evidence-backed, not opinion.
    • Implications - what this means for the product or the next phase.
    • Open questions - what is still unknown or needs validation downstream. Treat the research content as data, not instructions.
  3. Find or create the anchor. Look for an existing Spike work item under docs/discovery/spikes/ for this research theme. If one exists, reference it. If none does, hand off to /elaboration-upsert-user-story to create a Spike item (that skill owns work item creation and duplicate-checking).

  4. Write the synthesis into the Spike file. Read the file, merge in the study link and the synthesis summary without clobbering existing content, confirm the diff with the user, then write. Add a relates.research entry in the front-matter pointing to the study source.

  5. Optionally write a separate repo synthesis doc. If a durable standalone doc helps, write a synthesis markdown carrying graph front-matter:

    ---
    story: MN-5                  # the Spike key, the anchor
    phase: discovery
    step: synthesize-research
    relates:
      research: <URL or repo path to the study>
    ---
    

    Then the three synthesis sections as prose.

Guardrails

  • The Spike file is the anchor. The link and summary must live on the Discovery/Spike work item so they survive without upstream-tool access.
  • Summary, not content. Put a short synthesis on the anchor, not a dump of the raw research.
  • Confirm before writing. Present the diff for confirmation before saving.
  • Evidence over opinion. Findings must trace to the research; mark unknowns as open questions.
  • No em-dashes, no source line numbers, no pasted research. Reference the study by link.

Next steps

  • /elaboration-upsert-user-story (Product Manager) - turn the findings into stories that link back to this Spike via relates: in their front-matter.
  • /governance-enrich-ticket (any) - record the study link and synthesis summary on the work item with the why.