[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M77MNBNE8WZBTVCD3F9X696H":3,"skill-same:01M77MNBNE8WZBTVCD3F9X696H":98},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":28,"commitSha":29,"skillDirectory":30,"forksCount":31,"repositoryHtmlUrl":32,"repositoryHomepageUrl":6,"ownerHtmlUrl":33,"ownerBlogUrl":34,"ownerDescription":35,"updatedAt":36,"install":37,"resources":49},"01M77MNBNE8WZBTVCD3F9X696H","retro",null,"Structured session retrospective: transcript metrics, quality across five dimensions, feedback-memory regression check, and codified learnings. Use when: 'retro', 'what did we learn', or at end of session. Mid-session: \u002Fsession-flow:running-retro.","melodic-software\u002Fclaude-code-plugins","melodic-software","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F58273638?v=4","01M77JVQTXNGKCE4MDFY2BC936","Melodic Software","01M77JVQVN9QDWKAH6W03ANP76",30,"MIT","https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fsession-flow\u002Fskills\u002Fretro\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\ndescription: \"Structured session retrospective: transcript metrics, quality across five dimensions, feedback-memory regression check, and codified learnings. Use when: 'retro', 'what did we learn', or at end of session. Mid-session: \u002Fsession-flow:running-retro.\"\nargument-hint: \"[session [--pattern '\u003Cregex>']|quick] [--unattended]\"\nuser-invocable: true\ndisable-model-invocation: false\nmetadata:\n  workflow-stage: retro\n  summary: Structured session retrospective with codified learnings\n---\n\n**Arguments.** `[session [--pattern '\u003Cregex>']|quick] [--unattended]`. e.g., \u002Fretro, \u002Fretro session, \u002Fretro session --pattern 'PROJ-\\d+', \u002Fretro quick, \u002Fretro quick --unattended\n\n## Context. Gather first\n\nTake `session-id`, `branch`, `status`, `recent-commits` at `-5`, and `changed-files`.\nProbe commands, the one-command-per-call and treat-failure-as-unknown rules, and the `$`-expansion\nrationale for gathering at run time rather than pre-computing:\n[`${CLAUDE_PLUGIN_ROOT}\u002Freference\u002Fgather.md`](${CLAUDE_PLUGIN_ROOT}\u002Freference\u002Fgather.md).\n\n## Purpose\n\nThe self-improvement loop. Answers: \"What happened, what did we learn, and how do we prevent the\nsame mistakes next time?\"\n\n**Three concepts this skill enforces:**\n\n1. **Analyze**. Examine what happened with evidence (transcript metrics, conversation context,\n   feedback regressions)\n2. **Identify**. Find errors, behavioral adjustments, process improvements, and skill\u002Ftool\n   candidates\n3. **Codify**. Persist learnings durably (the consuming repo's instruction files and rules, or\n   Claude Code auto-memory for contributor-specific facts)\n\n**What this skill is NOT:** not a code review (design judgment on current work), not outcome\nverification (does the change match intent), and not Claude Code's built-in `\u002Finsights`\n(cross-session usage analytics); this is structured quality analysis with codification.\n\n## Paths\n\nResolve at runtime, never hardcode machine-specific paths:\n\n- **Session data root**. `~\u002F.claude\u002Fprojects\u002F\u003Cproject-slug>\u002F`, where `\u003Cproject-slug>` is the\n  project's absolute path with every character outside `[A-Za-z0-9]` replaced by `-`:\n\n  ```bash\n  PROJECT_SLUG=$(pwd -W 2>\u002Fdev\u002Fnull || pwd)          # Windows drive form when available\n  PROJECT_SLUG=$(printf '%s' \"$PROJECT_SLUG\" | sed 's\u002F[^A-Za-z0-9]\u002F-\u002Fg')\n  SESSION_DATA_DIR=\"$HOME\u002F.claude\u002Fprojects\u002F$PROJECT_SLUG\"\n  ```\n\n  Verify the directory exists before use; if the computed slug misses, find it by locating the\n  current session's JSONL: `ls \"$HOME\u002F.claude\u002Fprojects\"\u002F*\u002F\"${CLAUDE_CODE_SESSION_ID}.jsonl\"`.\n- Transcript: `\u003CSESSION_DATA_DIR>\u002F\u003Csession-id>.jsonl`; subagents:\n  `\u003CSESSION_DATA_DIR>\u002F\u003Csession-id>\u002Fsubagents\u002F`\n- **Auto-memory** (feedback regression check): `\u003CSESSION_DATA_DIR>\u002Fmemory\u002F`. Present only when the\n  consumer uses Claude Code auto-memory; degrade gracefully when absent\n- **Score history** (plugin state): `${CLAUDE_PLUGIN_DATA}\u002Fscores\u002F\u003Cproject-slug>.md`. Survives\n  plugin updates, never lands in the consumer's repo\n- Parser: `${CLAUDE_PLUGIN_ROOT}\u002Fskills\u002Fretro\u002Fscripts\u002Fparse_transcript.py` (stdlib-only,\n  Python 3.10+)\n\n## Step 0: Detect mode\n\n| Signal | Mode | Context file |\n|--------|------|-------------|\n| End of session, bare `\u002Fsession-flow:retro`, post-merge | **session** | `context\u002Fsession.md`, full 5-phase analysis |\n| \"quick retro\", short session, limited context | **quick** | `context\u002Fquick.md`, abbreviated pass |\n\nScore history across sessions names no retro mode. Before reading any mode, stop on a bare \"how am\nI doing\", a request for scores, or a `trends` argument, and name `\u002Fsession-flow:audit-sessions trends`;\n\"how did I do this session\", a question about this one session, still runs a retro.\nThat action reads the history this skill writes and adds delivery metrics; this skill analyzes no\nsession for it.\n\nA retro of one pull request (its sessions, its review comments) names no retro mode: stop and name\n`\u002Fsession-flow:audit-sessions pr --number \u003Cn> --repo '\u003Cowner\u002Frepo>'`.\n\nA `codify` argument, \"codify reviews\" or \"mine review comments\" names no retro mode: one recurring\nmistake, or the lessons in recent merged pull requests' review comments, go to\n`\u002Fcorrections:codify` (its `reviews` action reads the review comments) when the `corrections`\nplugin is installed. Name that skill and stop; when it is not installed, say that no installed\nskill codifies one mistake or mines review comments, and stop.\n\n`--unattended`, anywhere in `$ARGUMENTS`, is the caller declaring that nobody is present: take it\nout first. It only gates the user-only built-ins (`\u002Fexport`, `\u002Finsights`): an unattended run neither suggests nor runs them. The bare\nword `unattended` is neither this flag nor a mode: stop and name `--unattended`. A `--pattern '\u003Cregex>'` pair is taken out next and\nkept for session mode's ticket section; it selects `session` when no mode is named, and with\nanother mode named, stop and say `--pattern` is session mode's only. Read the mode from the remainder. If it specifies a\nmode, use it. Otherwise infer from context; when the session is long or\ndegraded, or compaction has occurred, prefer `quick`; ambiguous → `session`. Read the mode's context\nfile before proceeding.\n\n## Step 1: Execute the mode\n\nFollow the selected context file. Each mode has its own phases, outputs, and interactive\ncheckpoints.\n\n### Reference index. Load on demand\n\n| File | Load when |\n|---|---|\n| [reference\u002Fecosystem-improvement-catalog.md](reference\u002Fecosystem-improvement-catalog.md) | Before filling session mode's Phase 3 recommendation table, and any other time a finding has to be mapped to an ecosystem target. It owns the project-vs-personal placement decision tree and the per-target recommendation format for memory, rules, hooks, skills, agents, MCP servers, and settings. |\n\n### Session settings\n\nThe user option session mode reads, as rendered for this run (a literal, unexpanded placeholder\nmeans unset): `retro_lenses` = `${user_config.retro_lenses}`. `context\u002Fsession.md` \"Review lenses\"\nresolves it against the repository's `docs\u002Fconventions\u002Fsession-flow.yaml`, which wins.\n\n## Step 2: Handoff\n\nAfter the retrospective:\n\n| Condition | Suggestion |\n|-----------|-----------|\n| End-of-session, retro complete | Suggest any wrap-up steps the consuming repo defines |\n| Session mode, follow-ups queued | Suggest filing them in the consumer's work-item tracker |\n\n## Multi-session awareness\n\nWhen the sibling `handoff` skill's save-points exist (the resolved `\u003Cmemory_dir>\u002Fhandoffs\u002F`,\ndefault `.work\u002Fhandoffs\u002F`; or the consuming repo's documented location), the retro spans the\nwhole session CHAIN, not just the current session: the\nparser's `--chain-from` walks `previous_handoff` frontmatter pointers\nbackwards from the newest handoff file and aggregates metrics across every chained transcript. See\n`context\u002Fsession.md` Phase 1.\n\n**State the discovery basis, and never present a low-coverage chain retro silently.** The walk\nfollows `previous_handoff` pointers, so it stops at the first session that wrote no handoff file,\na chain linked by hand-pasted continuation prompts instead of save-points can end after one hop.\nThe parser reports what it saw in `chain_coverage` (`requested` \u002F `found` \u002F `available` \u002F `ratio`,\nwhere `available` is the requested chain, plus under `--chain-from` the other transcripts that\nmention the handoff's topic; `project_transcripts` counts the whole project directory apart from\nthe ratio). When `ratio` is below ~0.5, say\nso before presenting: name the found and available counts, and offer `--sessions` with the ids\nenumerated explicitly. A retro authored from a fifth of the evidence must not read like a complete\none. Offer each `fork_candidates` entry the same way: a fork shares records with the chain but no\nhandoff points at it, so the walk never reaches it.\n\n**Offer a durable copy when the session was worth retrospecting.** The transcripts this retro reads\nare retention-swept (`cleanupPeriodDays`), and the conversation itself has no durable artifact.\n\n- **Pointer**: [Data usage § Data retention](https:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fdata-usage#data-retention),\n  for the retention period and what `cleanupPeriodDays` adjusts.\n\nIf \u002Fexport is available in your session (gate basis: **Verification record: `\u002Fexport`** below), suggest that the person run it for a durable conversation copy at `\u003Cmemory_dir>\u002Fexports\u002F\u003CYYYYMMDDTHHMMSSZ>-\u003Ctopic>.txt` after verifying the memory root's self-ignore guard (a `.gitignore` containing `*`; create it and announce it when absent). This skill never invokes `\u002Fexport` itself. **`--unattended`:** make no suggestion.\n\n## Verification record: `\u002Fexport`\n\n- **Decision.** Suggest `\u002Fexport` to the person and never run it: it is a built-in interactive command the Skill tool does not list and headless runs refuse. A command that is not available to the person is left out of the menu.\n- **Pointer.** The `\u002Fexport [filename]` row of [Commands § All commands](https:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fcommands#all-commands). Probe: `claude --bare -p \"\u002Fexport \u003Cpath>\"` refused on Claude Code 2.1.241 (2026-08-24); invocation mode local-jsx on 2.1.263 (2026-09-11).\n- **As of.** 2026-09-29.\n- **Recheck when.** A Claude Code release note or the commands page adds an `\u002Fexport` format or redaction flag, a headless or programmatic form, or an official conversation-sharing surface.\n\n## Boundary, the built-in `\u002Finsights` command\n\nBoth look back at how sessions went, so \"how did I do\" can land on either.\n\n- **`\u002Finsights` (built-in command)**: a cross-session usage report; this skill never runs it. What\n  it reports: the records in\n  [reference\u002Fnative-insights.md](reference\u002Fnative-insights.md).\n- **This skill (marketplace plugin).** A structured retrospective of one session or handoff chain:\n  transcript metrics, five quality dimensions, feedback-memory regressions, and codification into\n  rules or memory behind approval.\n\n**Routing.** When the person asks about patterns across many sessions, offer it to the person: If \u002Finsights is available in your session (gate basis: ungated, though\nsome hosts lack it; records in [reference\u002Fnative-insights.md](reference\u002Fnative-insights.md)), you\ncan run it instead of or alongside this skill for a cross-session usage report. Make the offer at the end of the run. Prefer this skill for what one\nsession taught and for codifying it. When the caller passes `--unattended`, the run makes no\noffer.\n\n**Mutation gate.** This skill writes only approved codifications. It never runs `\u002Finsights` on the\nperson's behalf.\n\n**Availability is never assumed.** This section states what to do when the person can run\n`\u002Finsights`, never that it is present. What the report covers and where it is unavailable are read\nlive from the records in [reference\u002Fnative-insights.md](reference\u002Fnative-insights.md).\n\n## What this skill does NOT do\n\n- **Does not run builds or tests**. That's the consuming repo's verify stage\n- **Does not review code quality**. That's its review stage\n- **Does not write scores or reports into the consumer's repo**. Plugin state stays in\n  `${CLAUDE_PLUGIN_DATA}`; only user-approved codifications (rule edits, memory entries) land\n  outside it\n- **Does not read a consumer-supplied scoring rubric**. The five dimensions are fixed plugin\n  identity, not consumer config, and there is no seam to swap them. What adapts is what each\n  dimension scores *against* (your repo's conventions, session-type calibration), never the\n  dimensions themselves.\n\n## Spoke paths\n\nThe `context\u002F` files write the plugin's root directory as `\u003Cplugin-root>`, which is\n`${CLAUDE_PLUGIN_ROOT}`. Put that path in place of the placeholder before running a command or\nwriting it into a brief: a `${…}` token in a spoke reaches the Bash tool unsubstituted.\n\n- **Pointer**: \u003Chttps:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fplugins\u002Fmanifest-reference#where-each-variable-resolves>\n\n## Next\n\n- \u002Fsession-flow:audit-sessions trends: the score history session and quick modes write, across\n  sessions, with delivery metrics.\n- \u002Fsession-flow:audit-sessions sweep: checks whether what this retro found recurs across every\n  session on the machine.\n- \u002Fsession-flow:audit-sessions ticket --pattern '\u003Cregex>': every stored session for one work item.\n  Session mode calls it when given `--pattern` (`context\u002Fsession.md`, \"Ticket section\").\n\n## Gotchas\n\n- **Run all phases by default** in session mode. Skip metrics only when the parser fails or the\n  user asks\n- **Always include skill-candidate and follow-up-candidate analysis**, even when the conclusion is\n  \"no candidates this session\"\n- **Codifying follows the workflow too**. A finding that lands as instruction text (a `CLAUDE.md`,\n  `AGENTS.md` or rules-file line, the guidance tier of\n  [reference\u002Fenforcement-ladder.md](reference\u002Fenforcement-ladder.md)) names its proof in the PR:\n  an eval case, or a deletion watch with a date (`\u002Fharness-config:unhobble`, when the\n  `harness-config` plugin is installed). A mechanical finding goes to `\u002Fcorrections:codify` (when\n  the `corrections` plugin is installed), which proves its own check\n- **Phase 4 is an interactive checkpoint**, never persist codifications without explicit user\n  approval\n","**Arguments.** `[session [--pattern '\u003Cregex>']|quick] [--unattended]`. e.g., \u002Fretro, \u002Fretro session, \u002Fretro session --pattern 'PROJ-\\d+', \u002Fretro quick, \u002Fretro quick --unattended\n\n## Context. Gather first\n\nTake `session-id`, `branch`, `status`, `recent-commits` at `-5`, and `changed-files`.\nProbe commands, the one-command-per-call and treat-failure-as-unknown rules, and the `$`-expansion\nrationale for gathering at run time rather than pre-computing:\n[`${CLAUDE_PLUGIN_ROOT}\u002Freference\u002Fgather.md`](${CLAUDE_PLUGIN_ROOT}\u002Freference\u002Fgather.md).\n\n## Purpose\n\nThe self-improvement loop. Answers: \"What happened, what did we learn, and how do we prevent the\nsame mistakes next time?\"\n\n**Three concepts this skill enforces:**\n\n1. **Analyze**. Examine what happened with evidence (transcript metrics, conversation context,\n   feedback regressions)\n2. **Identify**. Find errors, behavioral adjustments, process improvements, and skill\u002Ftool\n   candidates\n3. **Codify**. Persist learnings durably (the consuming repo's instruction files and rules, or\n   Claude Code auto-memory for contributor-specific facts)\n\n**What this skill is NOT:** not a code review (design judgment on current work), not outcome\nverification (does the change match intent), and not Claude Code's built-in `\u002Finsights`\n(cross-session usage analytics); this is structured quality analysis with codification.\n\n## Paths\n\nResolve at runtime, never hardcode machine-specific paths:\n\n- **Session data root**. `~\u002F.claude\u002Fprojects\u002F\u003Cproject-slug>\u002F`, where `\u003Cproject-slug>` is the\n  project's absolute path with every character outside `[A-Za-z0-9]` replaced by `-`:\n\n  ```bash\n  PROJECT_SLUG=$(pwd -W 2>\u002Fdev\u002Fnull || pwd)          # Windows drive form when available\n  PROJECT_SLUG=$(printf '%s' \"$PROJECT_SLUG\" | sed 's\u002F[^A-Za-z0-9]\u002F-\u002Fg')\n  SESSION_DATA_DIR=\"$HOME\u002F.claude\u002Fprojects\u002F$PROJECT_SLUG\"\n  ```\n\n  Verify the directory exists before use; if the computed slug misses, find it by locating the\n  current session's JSONL: `ls \"$HOME\u002F.claude\u002Fprojects\"\u002F*\u002F\"${CLAUDE_CODE_SESSION_ID}.jsonl\"`.\n- Transcript: `\u003CSESSION_DATA_DIR>\u002F\u003Csession-id>.jsonl`; subagents:\n  `\u003CSESSION_DATA_DIR>\u002F\u003Csession-id>\u002Fsubagents\u002F`\n- **Auto-memory** (feedback regression check): `\u003CSESSION_DATA_DIR>\u002Fmemory\u002F`. Present only when the\n  consumer uses Claude Code auto-memory; degrade gracefully when absent\n- **Score history** (plugin state): `${CLAUDE_PLUGIN_DATA}\u002Fscores\u002F\u003Cproject-slug>.md`. Survives\n  plugin updates, never lands in the consumer's repo\n- Parser: `${CLAUDE_PLUGIN_ROOT}\u002Fskills\u002Fretro\u002Fscripts\u002Fparse_transcript.py` (stdlib-only,\n  Python 3.10+)\n\n## Step 0: Detect mode\n\n| Signal | Mode | Context file |\n|--------|------|-------------|\n| End of session, bare `\u002Fsession-flow:retro`, post-merge | **session** | `context\u002Fsession.md`, full 5-phase analysis |\n| \"quick retro\", short session, limited context | **quick** | `context\u002Fquick.md`, abbreviated pass |\n\nScore history across sessions names no retro mode. Before reading any mode, stop on a bare \"how am\nI doing\", a request for scores, or a `trends` argument, and name `\u002Fsession-flow:audit-sessions trends`;\n\"how did I do this session\", a question about this one session, still runs a retro.\nThat action reads the history this skill writes and adds delivery metrics; this skill analyzes no\nsession for it.\n\nA retro of one pull request (its sessions, its review comments) names no retro mode: stop and name\n`\u002Fsession-flow:audit-sessions pr --number \u003Cn> --repo '\u003Cowner\u002Frepo>'`.\n\nA `codify` argument, \"codify reviews\" or \"mine review comments\" names no retro mode: one recurring\nmistake, or the lessons in recent merged pull requests' review comments, go to\n`\u002Fcorrections:codify` (its `reviews` action reads the review comments) when the `corrections`\nplugin is installed. Name that skill and stop; when it is not installed, say that no installed\nskill codifies one mistake or mines review comments, and stop.\n\n`--unattended`, anywhere in `$ARGUMENTS`, is the caller declaring that nobody is present: take it\nout first. It only gates the user-only built-ins (`\u002Fexport`, `\u002Finsights`): an unattended run neither suggests nor runs them. The bare\nword `unattended` is neither this flag nor a mode: stop and name `--unattended`. A `--pattern '\u003Cregex>'` pair is taken out next and\nkept for session mode's ticket section; it selects `session` when no mode is named, and with\nanother mode named, stop and say `--pattern` is session mode's only. Read the mode from the remainder. If it specifies a\nmode, use it. Otherwise infer from context; when the session is long or\ndegraded, or compaction has occurred, prefer `quick`; ambiguous → `session`. Read the mode's context\nfile before proceeding.\n\n## Step 1: Execute the mode\n\nFollow the selected context file. Each mode has its own phases, outputs, and interactive\ncheckpoints.\n\n### Reference index. Load on demand\n\n| File | Load when |\n|---|---|\n| [reference\u002Fecosystem-improvement-catalog.md](reference\u002Fecosystem-improvement-catalog.md) | Before filling session mode's Phase 3 recommendation table, and any other time a finding has to be mapped to an ecosystem target. It owns the project-vs-personal placement decision tree and the per-target recommendation format for memory, rules, hooks, skills, agents, MCP servers, and settings. |\n\n### Session settings\n\nThe user option session mode reads, as rendered for this run (a literal, unexpanded placeholder\nmeans unset): `retro_lenses` = `${user_config.retro_lenses}`. `context\u002Fsession.md` \"Review lenses\"\nresolves it against the repository's `docs\u002Fconventions\u002Fsession-flow.yaml`, which wins.\n\n## Step 2: Handoff\n\nAfter the retrospective:\n\n| Condition | Suggestion |\n|-----------|-----------|\n| End-of-session, retro complete | Suggest any wrap-up steps the consuming repo defines |\n| Session mode, follow-ups queued | Suggest filing them in the consumer's work-item tracker |\n\n## Multi-session awareness\n\nWhen the sibling `handoff` skill's save-points exist (the resolved `\u003Cmemory_dir>\u002Fhandoffs\u002F`,\ndefault `.work\u002Fhandoffs\u002F`; or the consuming repo's documented location), the retro spans the\nwhole session CHAIN, not just the current session: the\nparser's `--chain-from` walks `previous_handoff` frontmatter pointers\nbackwards from the newest handoff file and aggregates metrics across every chained transcript. See\n`context\u002Fsession.md` Phase 1.\n\n**State the discovery basis, and never present a low-coverage chain retro silently.** The walk\nfollows `previous_handoff` pointers, so it stops at the first session that wrote no handoff file,\na chain linked by hand-pasted continuation prompts instead of save-points can end after one hop.\nThe parser reports what it saw in `chain_coverage` (`requested` \u002F `found` \u002F `available` \u002F `ratio`,\nwhere `available` is the requested chain, plus under `--chain-from` the other transcripts that\nmention the handoff's topic; `project_transcripts` counts the whole project directory apart from\nthe ratio). When `ratio` is below ~0.5, say\nso before presenting: name the found and available counts, and offer `--sessions` with the ids\nenumerated explicitly. A retro authored from a fifth of the evidence must not read like a complete\none. Offer each `fork_candidates` entry the same way: a fork shares records with the chain but no\nhandoff points at it, so the walk never reaches it.\n\n**Offer a durable copy when the session was worth retrospecting.** The transcripts this retro reads\nare retention-swept (`cleanupPeriodDays`), and the conversation itself has no durable artifact.\n\n- **Pointer**: [Data usage § Data retention](https:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fdata-usage#data-retention),\n  for the retention period and what `cleanupPeriodDays` adjusts.\n\nIf \u002Fexport is available in your session (gate basis: **Verification record: `\u002Fexport`** below), suggest that the person run it for a durable conversation copy at `\u003Cmemory_dir>\u002Fexports\u002F\u003CYYYYMMDDTHHMMSSZ>-\u003Ctopic>.txt` after verifying the memory root's self-ignore guard (a `.gitignore` containing `*`; create it and announce it when absent). This skill never invokes `\u002Fexport` itself. **`--unattended`:** make no suggestion.\n\n## Verification record: `\u002Fexport`\n\n- **Decision.** Suggest `\u002Fexport` to the person and never run it: it is a built-in interactive command the Skill tool does not list and headless runs refuse. A command that is not available to the person is left out of the menu.\n- **Pointer.** The `\u002Fexport [filename]` row of [Commands § All commands](https:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fcommands#all-commands). Probe: `claude --bare -p \"\u002Fexport \u003Cpath>\"` refused on Claude Code 2.1.241 (2026-08-24); invocation mode local-jsx on 2.1.263 (2026-09-11).\n- **As of.** 2026-09-29.\n- **Recheck when.** A Claude Code release note or the commands page adds an `\u002Fexport` format or redaction flag, a headless or programmatic form, or an official conversation-sharing surface.\n\n## Boundary, the built-in `\u002Finsights` command\n\nBoth look back at how sessions went, so \"how did I do\" can land on either.\n\n- **`\u002Finsights` (built-in command)**: a cross-session usage report; this skill never runs it. What\n  it reports: the records in\n  [reference\u002Fnative-insights.md](reference\u002Fnative-insights.md).\n- **This skill (marketplace plugin).** A structured retrospective of one session or handoff chain:\n  transcript metrics, five quality dimensions, feedback-memory regressions, and codification into\n  rules or memory behind approval.\n\n**Routing.** When the person asks about patterns across many sessions, offer it to the person: If \u002Finsights is available in your session (gate basis: ungated, though\nsome hosts lack it; records in [reference\u002Fnative-insights.md](reference\u002Fnative-insights.md)), you\ncan run it instead of or alongside this skill for a cross-session usage report. Make the offer at the end of the run. Prefer this skill for what one\nsession taught and for codifying it. When the caller passes `--unattended`, the run makes no\noffer.\n\n**Mutation gate.** This skill writes only approved codifications. It never runs `\u002Finsights` on the\nperson's behalf.\n\n**Availability is never assumed.** This section states what to do when the person can run\n`\u002Finsights`, never that it is present. What the report covers and where it is unavailable are read\nlive from the records in [reference\u002Fnative-insights.md](reference\u002Fnative-insights.md).\n\n## What this skill does NOT do\n\n- **Does not run builds or tests**. That's the consuming repo's verify stage\n- **Does not review code quality**. That's its review stage\n- **Does not write scores or reports into the consumer's repo**. Plugin state stays in\n  `${CLAUDE_PLUGIN_DATA}`; only user-approved codifications (rule edits, memory entries) land\n  outside it\n- **Does not read a consumer-supplied scoring rubric**. The five dimensions are fixed plugin\n  identity, not consumer config, and there is no seam to swap them. What adapts is what each\n  dimension scores *against* (your repo's conventions, session-type calibration), never the\n  dimensions themselves.\n\n## Spoke paths\n\nThe `context\u002F` files write the plugin's root directory as `\u003Cplugin-root>`, which is\n`${CLAUDE_PLUGIN_ROOT}`. Put that path in place of the placeholder before running a command or\nwriting it into a brief: a `${…}` token in a spoke reaches the Bash tool unsubstituted.\n\n- **Pointer**: \u003Chttps:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fplugins\u002Fmanifest-reference#where-each-variable-resolves>\n\n## Next\n\n- \u002Fsession-flow:audit-sessions trends: the score history session and quick modes write, across\n  sessions, with delivery metrics.\n- \u002Fsession-flow:audit-sessions sweep: checks whether what this retro found recurs across every\n  session on the machine.\n- \u002Fsession-flow:audit-sessions ticket --pattern '\u003Cregex>': every stored session for one work item.\n  Session mode calls it when given `--pattern` (`context\u002Fsession.md`, \"Ticket section\").\n\n## Gotchas\n\n- **Run all phases by default** in session mode. Skip metrics only when the parser fails or the\n  user asks\n- **Always include skill-candidate and follow-up-candidate analysis**, even when the conclusion is\n  \"no candidates this session\"\n- **Codifying follows the workflow too**. A finding that lands as instruction text (a `CLAUDE.md`,\n  `AGENTS.md` or rules-file line, the guidance tier of\n  [reference\u002Fenforcement-ladder.md](reference\u002Fenforcement-ladder.md)) names its proof in the PR:\n  an eval case, or a deletion watch with a date (`\u002Fharness-config:unhobble`, when the\n  `harness-config` plugin is installed). A mechanical finding goes to `\u002Fcorrections:codify` (when\n  the `corrections` plugin is installed), which proves its own check\n- **Phase 4 is an interactive checkpoint**, never persist codifications without explicit user\n  approval\n",{"metadata":24,"description":7,"argument-hint":26,"user-invocable":27,"disable-model-invocation":20},{"summary":25,"workflow-stage":5},"Structured session retrospective with codified learnings","[session [--pattern '\u003Cregex>']|quick] [--unattended]",true,"main","4f2dcaaff35e8f9fccc6a9f4d82a5830ab17a5c5","plugins\u002Fsession-flow\u002Fskills\u002Fretro",7,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins","https:\u002F\u002Fgithub.com\u002Fmelodic-software","https:\u002F\u002Fmelodicsoftware.com","Independent software studio building developer tools, automation, and application services, with roots in music.","2026-11-15 05:15:06",{"source":8,"skillName":5,"commands":38,"prompt":48,"zipAvailable":27},[39,42,45],{"tool":40,"command":41},"npx","npx skills add melodic-software\u002Fclaude-code-plugins --skill retro",{"tool":43,"command":44},"bunx","bunx skills add melodic-software\u002Fclaude-code-plugins --skill retro",{"tool":46,"command":47},"pnpm","pnpm dlx skills add melodic-software\u002Fclaude-code-plugins --skill retro","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fsession-flow\u002Fskills\u002Fretro\u002FSKILL.md",[50,56,60,64,68,72,76,82,86,90,94],{"path":51,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":54,"sourceHtmlUrl":55},"context\u002Fquick.md","REFERENCE","参考文档",2299,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fsession-flow\u002Fskills\u002Fretro\u002Fcontext\u002Fquick.md",{"path":57,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":58,"sourceHtmlUrl":59},"context\u002Fsession.md",27514,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fsession-flow\u002Fskills\u002Fretro\u002Fcontext\u002Fsession.md",{"path":61,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":62,"sourceHtmlUrl":63},"evals\u002Fevals.json",12494,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fsession-flow\u002Fskills\u002Fretro\u002Fevals\u002Fevals.json",{"path":65,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":66,"sourceHtmlUrl":67},"reference\u002Fecosystem-improvement-catalog.md",12366,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fsession-flow\u002Fskills\u002Fretro\u002Freference\u002Fecosystem-improvement-catalog.md",{"path":69,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":70,"sourceHtmlUrl":71},"reference\u002Fenforcement-ladder.md",13652,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fsession-flow\u002Fskills\u002Fretro\u002Freference\u002Fenforcement-ladder.md",{"path":73,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":74,"sourceHtmlUrl":75},"reference\u002Fnative-insights.md",1638,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fsession-flow\u002Fskills\u002Fretro\u002Freference\u002Fnative-insights.md",{"path":77,"resourceType":78,"resourceTypeDesc":79,"sizeBytes":80,"sourceHtmlUrl":81},"scripts\u002Fparse_transcript.py","SCRIPT","脚本",35656,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fsession-flow\u002Fskills\u002Fretro\u002Fscripts\u002Fparse_transcript.py",{"path":83,"resourceType":78,"resourceTypeDesc":79,"sizeBytes":84,"sourceHtmlUrl":85},"scripts\u002Fparse-concern-value.sh",5082,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fsession-flow\u002Fskills\u002Fretro\u002Fscripts\u002Fparse-concern-value.sh",{"path":87,"resourceType":78,"resourceTypeDesc":79,"sizeBytes":88,"sourceHtmlUrl":89},"scripts\u002Fparse-transcript.test.sh",745,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fsession-flow\u002Fskills\u002Fretro\u002Fscripts\u002Fparse-transcript.test.sh",{"path":91,"resourceType":78,"resourceTypeDesc":79,"sizeBytes":92,"sourceHtmlUrl":93},"scripts\u002Ftest_parse_transcript.py",37810,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fsession-flow\u002Fskills\u002Fretro\u002Fscripts\u002Ftest_parse_transcript.py",{"path":95,"resourceType":78,"resourceTypeDesc":79,"sizeBytes":96,"sourceHtmlUrl":97},"scripts\u002Fyaml-subset.awk",9901,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fsession-flow\u002Fskills\u002Fretro\u002Fscripts\u002Fyaml-subset.awk",[99,106,113,120,125,132],{"skillNo":100,"name":101,"nameZh":6,"description":102,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":103,"categories":104,"keywords":105,"publishedAt":19},"01M7DYYDG31B9YX7N9MB9QW2GW","track","Backlog reads through the bound tracker: search open and closed items, view one item with its comments and state history. Use when another skill searches or reads tracker items.","https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fsession-flow\u002Fevals\u002Frecall-still-reports-reopened-bug\u002Fwork-items-stub\u002Fskills\u002Ftrack\u002FSKILL.md",[],[],{"skillNo":107,"name":108,"nameZh":6,"description":109,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":110,"categories":111,"keywords":112,"publishedAt":19},"01M7DYXVBAVG3NV30PA7GKW9Q6","reviewer-media","Reviewer media for one pull request: decides whether it needs a narrated walkthrough, builds the storyboard from \u002Freview:explain-change's digest, has \u002Fexplainer-video:produce render it and hands the manifest to \u002Fsource-control:pull-request media. Use when asked for 'reviewer media', 'a walkthrough video for this PR', 'a produced video of this change', or when pull-request ready or monitor hands a PR over.","https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Freview\u002Fskills\u002Freviewer-media\u002FSKILL.md",[],[],{"skillNo":114,"name":115,"nameZh":6,"description":116,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":117,"categories":118,"keywords":119,"publishedAt":19},"01M7DYRZ5YD44AMYH26YR25GZ8","route","Model and effort per agent role (orchestrator, worker, verifier, retrieval) for a multi-agent run, resolved through the bundled defaults and the user, team and personal override config layers into fenced JSON with single-agent and fan-out variants and the layer behind each value. Use when asked 'which model for the subagents', 'route the roles' or 'model routing', or before a skill passes a role's model on a spawn.","https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fdiscovery\u002Fevals\u002Fex-model-pin-without-override\u002Fmulti-agent-stub\u002Fskills\u002Froute\u002FSKILL.md",[],[],{"skillNo":121,"name":115,"nameZh":6,"description":116,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":122,"categories":123,"keywords":124,"publishedAt":19},"01M7DYRTVT4FWTGDJY0HCCFJMZ","https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fdiscovery\u002Fevals\u002Fex-model-override-from-route\u002Fmulti-agent-stub\u002Fskills\u002Froute\u002FSKILL.md",[],[],{"skillNo":126,"name":127,"nameZh":6,"description":128,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":129,"categories":130,"keywords":131,"publishedAt":19},"01M7DYNZC8PNW3T27V2ADKECKJ","principles","Module-design vocabulary for architecture findings, and the checks built on it. Use when: 'what is a deep module', 'what is a seam', 'deep vs shallow module', 'what is the deletion test', 'leverage and locality', 'module design vocabulary', 'which word do we use for X in architecture findings'. Scanning a codebase for friction is \u002Farchitecture:improve.","https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Farchitecture\u002Fskills\u002Fprinciples\u002FSKILL.md",[],[],{"skillNo":133,"name":134,"nameZh":6,"description":135,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":136,"categories":137,"keywords":138,"publishedAt":19},"01M7DYNMCZNEWBFHV9KTSS6H60","write","Structured five-field bug report (title, steps to reproduce, expected vs actual, justified severity, suggested fix location) from an informal description, read-only. Use when the user names a defect they observed, or asks for one written up ('report a bug', 'file a bug', 'write this up as a bug').","https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Farchitecture\u002Fevals\u002Fdeepening-defect-goes-to-bugs-write\u002Fbugs-stub\u002Fskills\u002Fwrite\u002FSKILL.md",[],[]]