Interaction retrospective
Skill iplaylf2/coding-profile/profile/agent/skills/interaction-retrospective
Personal coding profile kit for reusable defaults and setup
npx -y skills add iplaylf2/coding-profile --skill interaction-retrospectiveAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
2 things to look at
- no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
- 1 stars1 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.
What its author says it does
Copied from the file, not written here
Use when the user asks to analyze or debrief an agent interaction to decide whether a lesson should become durable guidance for future agent behavior, determine its persistence scope or destination, or ground a requested durable-guidance change in interaction evidence. Do not use for ordinary task summaries, code review, or project postmortems unless the goal is to improve future agent behavior.
SKILL.md
4.8 KB, as published. Nobody here has run it
Interaction Retrospective
Retrospective Boundary
This skill turns an interaction record into a persistence decision. It determines whether a durable lesson exists and, if so, the portable behavior, supported reach, and owning guidance source. Treat explicit user choices and already-established targets as inputs. Supply the lesson and its evidence-backed requirements without deciding the target artifact's structure beyond its governing schema and conventions.
Interaction Evidence Questions
- What future agent behavior is the user trying to improve or preserve?
- Which stage needs to differ: activation, context gathering, judgment, action, validation, communication, or another identified stage?
- Which agent choices, user corrections or endorsements, and outcomes show what helped or hindered the result?
- Which instructions, files, conventions, tool outputs, or prior messages were available, and how did the observed behavior respond to them?
- Which parts of the interaction reflect a legitimate choice, changing scope, subjective judgment, or context that only the user could supply, and which reveal a reusable behavior gap?
- What did convergence or success reveal: a portable practice or only a solution specific to this interaction?
Persistence Questions
- What has the user or an already-established task decision fixed about persistence, scope, or destination?
- Would durable guidance improve relevant future behavior, or is the lesson a one-off event, too uncertain, or already governed adequately?
- If guidance already expresses the desired behavior, was the problem its application rather than a missing or inaccurate rule?
- What interaction evidence or explicit persistence request supports the lesson's recurrence and reach, and which missing answer would materially change that judgment enough to require clarification?
- Which classification best fits the record: reasonable collaboration, an activation or execution gap, missing project context, a stable working preference, no durable lesson, or another supported explanation?
Scope And Destination Questions
- Is the durable content an agent reasoning or workflow rule, project-specific knowledge or process, a user working preference, or separate lessons of more than one kind?
- What reach does the record support: one project, a recurring task family, cross-project work for this user, or another explicit boundary?
- Which existing guidance source owns that kind of content and reach, or what distinct responsibility would justify a new source?
- If reusable agent guidance does not fit an existing task family, does the lesson establish a distinct recurring family with its own activation boundary, information sources, judgments, and validation responsibility that justifies a new skill?
- Would the selected destination duplicate, contradict, or overload an existing source of truth rather than improve it?
Guidance Requirement Questions
- What portable behavior should the selected guidance require or preserve?
- Under what conditions should that behavior apply, and what evidence limits its scope?
- What representative success or failure situation must the guidance help a future agent recognize and handle differently?
- Which private, incidental, or interaction-specific details explain this record but do not belong in durable guidance?
If no durable lesson or owning source is justified, leave durable guidance unchanged.
Decision Validation Questions
- Does the proposed lesson explain the original interaction and the observed improvement or failure?
- Would it guide supported analogous situations without changing unrelated behavior that the record does not cover?
- Does the selected scope and destination match the content they are responsible for owning?
- What validation was actually performed, and which conclusions remain static inferences or recommendations?
Result Questions
Lead with whether a durable lesson exists, including no change, then state its portable behavior, supported scope, and owning guidance source.
- What interaction evidence most strongly supports the decision?
- What guidance requirements follow from that evidence?
- What guidance change follows from the decision, and was it requested, recommended, or already applied?
- What was validated, and what uncertainty still materially affects persistence?