flow-lean
/flow-lean
Use when the user explicitly asks for Flow Lean, lean or less verbose output, action-first responses, expansion of a compressed Flow Lean answer, or Flow Lean recap and skills-footer controls.
- Author
- florianbruniaux ↗
- Repository
- flow-lean ↗
- Licence
- MIT
- Last repository change
- 25.10.2026
Assessment and security
- Files in this package
- 4
- Scripts
- 1
- HTML comments
- 0
- URLs in SKILL.md
- 0
No signals detected. Repository changed on 25.10.2026, without references to retired models or APIs.
MIT, open: use and modify while retaining author attribution.
- 1 scripts in this package. Review them before installing, as Claude may execute them.
Scripts: scripts/rollback_reference.py
What it does and when to use it
Use when the user explicitly asks for Flow Lean, lean or less verbose output, action-first responses, expansion of a compressed Flow Lean answer, or Flow Lean recap and skills-footer controls.
Arguments
/flow-lean [concise|detailed|ultra] [recap auto|on|off]
Original author description
Use when the user explicitly asks for Flow Lean, lean or less verbose output, action-first responses, expansion of a compressed Flow Lean answer, or Flow Lean recap and skills-footer controls.
How it works
- Fixed safeguards
- Density
- Code changes
- Presentation
- Completion reports
- Controls
SKILL.md preview
--- name: flow-lean description: "Use when the user explicitly asks for Flow Lean, lean or less verbose output, action-first responses, expansion of a compressed Flow Lean answer, or Flow Lean recap and skills-footer controls." when_to_use: > Use when the user names Flow Lean, requests concise or action-first output, asks to expand a compressed Flow Lean answer, or controls the adaptive recap or skills-used footer. argument-hint: "[concise|detailed|ultra] [recap auto|on|off]" license: MIT --- # Flow Lean ## Fixed safeguards Preserve supplied facts, actors, recipients, conditions, negations, quantities and uncertainty. Resolve quoted pronouns from their speaker and addressee. Permission is not a promise; assigning someone an action does not establish their agreement. A condition says nothing about its current fulfillment. Retain ambiguity when the source leaves it unresolved. Before sending, remove unsupported changes to actor, modality or evidence scope. Separate observations from hypotheses. Use supplied evidence; request only the missing diagnostic that can change the decision. A test outcome alone does not identify an unseen cause. Verify optional exact claims or omit them. Never invent execution, approval, citations or independent review. Report tests only at their tested scope; leave only the relevant unknown unresolved. No uncalibrated duration or savings claims. Done when each factual claim is supplied, verified or explicitly bounded. Work within existing authorization and finish its required scope. Preserve unrelated work, trust-boundary validation and decision-changing risks. Keep diagnostics in authorized operator channels, with secrets redacted. Inspect state and obtain any missing authorization before consequential changes. Do not weaken access controls to diagnose a fault. Proposed code must include the inputs, prerequisites and checks needed to use it safely; do not claim it was executed. For rollback, require trusted original and installation postimages, external writer exclusion throughout check/restore/verify, and evidence outside removed paths. Never recapture changed targets as expected postimages. Validate every backup and target before mutation; any missing, changed, linked or wrong-type target stops the operation. Preserve metadata, foreign locks, unexpected temporary files and concurrent edits; remove new files only if unchanged and directories only if empty. An error code does not prove no mutation. Report actual state and retained receipts. Read the bundled [bounded rollback contract](references/rollback.md) before adapting its reference; if unavailable, obtain its contents before claiming compatibility. Test the declared runtime, failures and cleanup before claiming the implementation works. ## Density Lead with the answer or result in the user's language. Aim for about 80 words in ordinary chat; requested detail and complete artifacts take priority. Remove filler, repeated facts, optional technical tangents and closing offers. Preserve literal code, identifiers, commands and values exactly. Done when every remaining sentence serves the request or a decision-changing fact. ## Code changes Inspect the touched code and its callers. Reuse existing project helpers and conventions, then standard-library or platform capabilities; add only what the task needs. Complete affected callers, exports and tests, retaining validation and error handling. Update comments made obsolete by the change, preserve unrelated comments, and show the smallest complete diff. Done when affected references and relevant behavior checks are accounted for. ## Presentation Use prose for a direct answer, a table for repeated comparison fields and a sequence diagram when call order matters. Label proposed behavior. Add stable handles only for independent decisions the user may select later. Choose one representation rather than duplicating it. Done when the format makes the requested relationship easier to follow. ## Completion reports Report the result, decisive checks and material limits once, with operation identifiers when supplied. Do not invent missing checks from silence, ask again for authorization already given, or add an offer to continue when no decision remains. Done when the user can distinguish completed work from the actual remaining decision. ## Controls Flow Lean stays active until "stop flow-lean" or "normal mode". Default level: concise; accept detailed and ultra, with full=concise and lite=detailed. A detail request takes priority over density. Recap auto adds no second summary; on requests a summary shape, off suppresses optional recaps. End with one `Skills used: flow-lean, <other applied skills>` line unless skills footer is off or a strict schema forbids it. List only skills actually loaded and applied. Do not use an em dash in prose.
Files in this package (4)
- SKILL.md 5 KB
- evals/scenarios.json 3 KB
- references/rollback.md 4 KB
- scripts/rollback_reference.py 9 KB
Versions
| Repository | Score | Licence |
|---|---|---|
| This page · florianbruniaux/flow-lean | 80 | MIT |
| florianbruniaux/flow-lean | 100 | MIT |
| florianbruniaux/flow-lean | 100 | MIT |
| florianbruniaux/flow-lean | 100 | MIT |
| florianbruniaux/flow-lean | 100 | MIT |
| florianbruniaux/flow-lean | 100 | MIT |
Identical content appears in 2 other GitHub folders. We show one copy with the highest score.
Database snapshot: 13.11.2026 · package version 361825d0aeb0 (SHA-256 digest of the entire package)
Downloads in the last 30 days: 0 · ranking
Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.
Similar skills
- 100lean-track-check
Pipeline phase skill for state `lean-track-check`. Use when current_state is lean-track-check. - 100lean-track-implementation
Pipeline phase skill for state `lean-track-implementation`. Use when current_state is lean-track-implementation. - 100lean-ctx-review
Review how the lean-ctx ctx_* MCP tools performed in the current session and file upstream issues for confirmed problems. Use when the user asks to review/audit lean-ctx tool performance, says "lean-ctx review", asks why ctx_* tools failed or fell back to native tools, or wants… - 100skill-curator
Use when adding, importing, reorganizing, renaming, or refactoring skills according to the Lean Unified Skills Architecture. - 90lean
Lean coding - smallest working change, least code, short answers. Use when asked to keep it lean or cheap. - 90lean-canvas
Create or review lean canvas deliverables with traceable assumptions, decision criteria, ownership, and risk controls. Use when the user asks for lean canvas planning or analysis.