agentsclimarketplace
Sign in

Retrospect domain

Skill pantheon-org/tekhne/skills/specialized/retrospect-domain

Analyze domain insights (WHAT/WHY learned) from captured sessions. Use when reviewing learnings, extracting patterns, analyzing decisions. Triggers include "retrospect domain", "domain analysis", "what did I learn", "session insights".From its SKILL.md

Install
npx -y skills add pantheon-org/tekhne --skill retrospect-domain

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 9 stars9 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

5.6 KB, ~1.2k tokens by cl100k_base, as published. Nobody here has run it

Retrospect Domain — Domain Insights Analysis

Analyze captured sessions to extract domain learnings: what worked, decisions made, patterns emerged.

Arguments

  • Domain (optional first arg): technical | research | generic (default)
  • Timeframe: --last Nd | --week | --month | --from DATE --to DATE | all

When to Use

Apply this skill when the user asks about:

  • What was learned in sessions ("what did I learn?", "session insights")
  • Extracting domain patterns or recurring decisions from a time window
  • Surfacing anti-patterns or gotchas discovered during work
  • Generating a domain knowledge summary across multiple sessions

When Not to Use

  • Use retrospect-collab instead when the focus is on HOW the collaboration worked, not WHAT was learned.
  • Do not run this skill against zero session files; no sessions means no domain insights to extract.

Principles

  • Extract, don't invent: Base every insight on actual session content. NEVER fabricate learnings not present in the session data.
  • Be specific: Reference actual conversations, tools used, and decisions made. ALWAYS name the session when citing evidence.
  • Detect patterns: Look for themes across multiple sessions, not isolated observations.
  • Domain-appropriate: Apply the domain framework matching the session type. Consider falling back to the generic framework only when domain content is ambiguous.
  • Framework flexibility: Optionally adapt analysis headings when session content does not fit the standard framework questions.

Anti-Patterns

<!-- BAD: "you learned to avoid N+1 queries" (invented) | GOOD: "session 2026-03-14 revealed an N+1 query in the reporting module — query was rewritten using eager loading" -->
  • NEVER report a learning that is not evidenced in the session — WHY: fabricated insights give false confidence and may direct future work incorrectly
  • NEVER apply the generic framework to clearly technical sessions — WHY: the Technical Domain framework surfaces six specific questions that generic analysis misses, reducing insight quality
  • NEVER omit the Patterns Emerged section when reviewing more than two sessions — WHY: cross-session synthesis is the primary value-add of this skill; single-session summaries are not enough
  • NEVER hardcode the output path — WHY: the .retro/insights/ location is project-configurable; use the configured path, not a hardcoded one

Steps

  1. Load domain framework (if specified):

    • Check for a domain-{domain}.md file at the project-configurable domain frameworks location.
    • If missing: warn, use generic framework

    Note: The domain frameworks location is project-configurable — point it to wherever your project stores domain reference files.

  2. Filter sessions:

    bash ${CLAUDE_PLUGIN_ROOT}/scripts/retrospect-load-sessions.sh $@

    First line: PERIOD: YYYY-MM-DD_to_YYYY-MM-DD. Remaining: session paths.

    Note: CLAUDE_PLUGIN_ROOT points to the script bundle location. The learnings output path (.retro/insights/) is project-configurable — set it to wherever your project stores retrospective outputs.

  3. Read session files — extract conversation turns, tools used, git context

  4. Analyze domain insights using framework:

    • What new patterns/approaches discovered?
    • What decisions were made and why?
    • What worked well / what failed?
    • What patterns recur across sessions?

  5. Generate Start/Stop/Continue recommendations

  6. Write insights to .retro/insights/domain/{PERIOD}.md

    # Output path example (project-configurable)
.retro/insights/domain/2026-03-01_to_2026-03-31.md

  7. Report: file path, suggest /retrospect collab or /retrospect report

Usage Examples

Run for the last 7 days with the technical domain framework:

/retrospect domain technical --last 7d

Run for the current month with no domain filter (generic):

/retrospect domain --month

Run for a custom date range with the research framework:

/retrospect domain research --from 2026-03-01 --to 2026-03-31

Run and then follow up with collaboration analysis:

/retrospect domain technical --week
/retrospect collab --week

Report Section Format

The domain report follows this structure:

# Domain Retrospective: YYYY-MM-DD to YYYY-MM-DD

## Sessions Reviewed
- [List sessions with timestamps and brief summaries]

## Domain Learnings
### What I Learned
### What Worked Well
### What Didn't Work
### Decisions Made
### Patterns Emerged

## Start/Stop/Continue
## Action Items
## Metrics Summary

Gotchas

  • The session filter script outputs a PERIOD header on line one; parse it separately from session file paths.
  • Domain framework files may not exist for every domain. ALWAYS warn the user and fall back to generic rather than aborting.
  • Consider including a short "Sessions Reviewed" list at the top of the report so the reader knows what data was used.
  • Run the session loader directly to test it: ./scripts/retrospect-load-sessions.sh --last 7d

References

  • Reference — scoring rubrics, metric definitions, and report format

What ships with it: 8 files

16.1 KB alongside SKILL.md

.tessl-plugin/

evals/

references/

Gives 0 of the 12 instructions most research analysis skills give in ~1.2k tokens

Counted across 1,063 of the 1,754 authors here whose files we hold, read 2026-08-07

  • Generate a markdown reportin 32 of 1063, across 23 files
  • Cite each claim's sourcein 30 of 1063, across 15 files
  • Define the ideal customer profilein 20 of 1063, across 2 files
  • Search for companies matching the criteriain 20 of 1063, across 2 files
  • Assign a fit score from one to tenin 20 of 1063, across 2 files
  • Analyze the codebase to understand the productin 19 of 1063, across 1 file
  • Ask clarifying questions about the value propositionin 19 of 1063, across 1 file
  • Look for signals of immediate needin 19 of 1063, across 1 file
  • Identify the target decision maker rolein 19 of 1063, across 1 file
  • Suggest a personalized contact strategyin 19 of 1063, across 1 file
  • Provide conversation starters for outreachin 19 of 1063, across 1 file
  • Format results in a scannable markdown templatein 19 of 1063, across 1 file

Said here and by no other author read

  • base every insight on actual session content
  • name the session when citing evidence
  • detect patterns across multiple sessions
  • load the matching domain framework
  • fall back to generic framework if domain file is missing
  • filter sessions using the load script

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.

Keep looking

Skills are one crate of 326,835. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.