laimis91/assistant-framework/docs/evals/variants/workflow-kernel-v1/SKILL.md
assistant-workflow
Run proportional development phases and resume persisted task state. Use to plan, build, implement, fix, migrate, refactor, or continue project artifacts.
- Source repository stars
- 7
- Declared platforms
- 0
- Static risk flags
- 0
- Last source update
- 2026-08-04
- Source checked
- 2026-08-04
Decision brief
What it does—and where it fits
Run proportional development phases and resume persisted task state. Use to plan, build, implement, fix, migrate, refactor, or continue project artifacts.
Not for
- Tasks that require unconfirmed production actions or broad system permissions.
- Environments where the pinned source and install steps cannot be inspected.
Compatibility matrix
Platform support, with evidence labels
| Platform | Status | Evidence | What to check |
|---|---|---|---|
| Codex | Not declared | No explicit evidence | Portability before use |
| Claude Code | Not declared | No explicit evidence | Portability before use |
| Cursor | Not declared | No explicit evidence | Portability before use |
| Gemini CLI | Not declared | No explicit evidence | Portability before use |
Installation
Inspect first. Install second.
The source command is displayed only when detected. A safe inspection prompt is always available so your agent can explain every action before execution.
npx skills add https://github.com/laimis91/assistant-framework --skill "docs/evals/variants/workflow-kernel-v1"Inspect the Agent Skill "assistant-workflow" from https://github.com/laimis91/assistant-framework/blob/e85fbadd83b022d9644bbe51b11784cf2e2dfd72/docs/evals/variants/workflow-kernel-v1/SKILL.md at commit e85fbadd83b022d9644bbe51b11784cf2e2dfd72. List every install step, command, network request, credential, file read/write, external action, and rollback step. Explain whether it fits my task. Do not install or execute anything until I approve.
Workflow
What the source asks the agent to do
- 01
Goal
Move project work from the newest accepted requirement to a verified, reviewed, reconstructable outcome with the least process that safely fits the risk.
Move project work from the newest accepted requirement to a verified, reviewed, reconstructable outcome with the least process that safely fits the risk. - 02
Success Criteria
Run the proportional phase path: Discover - optional Decompose - optional Plan -
Run the proportional phase path: Discover - optional Decompose - optional Plan -Reconcile persisted state against the newest user request and currentIf resume reconciliation classifies persisted state as stale, superseded, or completed, - 03
Constraints
Explicit user or repository artifact schemas override workflow-internal shapes; preserve exact paths, keys, types, ids, and supplied literals.
Explicit user or repository artifact schemas override workflow-internal shapes; preserve exact paths, keys, types, ids, and supplied literals.Preserve user changes, secrets, permissions, and company data boundaries.Prefer repository-native commands and patterns. Do not install or call - 04
Contract Loading
Read contracts/index.yaml first. Canonical contracts remain authoritative; load only the active selector:
entry for task inputs and triage;currentphase for the active transition and invariants;selectedhandoff before dispatch and return validation; - 05
Execution
1. Triage size, risk, gates, state, lane, review/QA, and delegation triggers. 2. Discover the relevant code and constraints; reconcile state. Create the Requirement Acceptance Map when its medium+/promotion condition applies. 3. Apply planmode: skip Plan for none, keep inline co…
Triage size, risk, gates, state, lane, review/QA, and delegation triggers.Discover the relevant code and constraints; reconcile state. Create theApply planmode: skip Plan for none, keep inline concise, and for
Permission review
Static risk signals and limitations
No configured static risk pattern was detected
This is not proof of safety. Runtime behavior, indirect dependencies, and hidden external systems are outside the static scan.
Evidence record
Why each signal appears
| Signal | Value | Evidence type | Meaning |
|---|---|---|---|
| Quality score | 90/100 | Computed | Documentation, specificity, maintenance, and trust rules |
| Repository stars | 7 | Source | Repository attention, not individual Skill quality |
| Compatibility | 0 platforms | Source | Declared in the catalog source record |
| Usage guide | automated source guide | Editorial | Generated or reviewed according to the visible evidence level |
Pinned source
Provenance and original SKILL.md
- Repository
- laimis91/assistant-framework
- Skill path
- docs/evals/variants/workflow-kernel-v1/SKILL.md
- Commit
- e85fbadd83b022d9644bbe51b11784cf2e2dfd72
- License
- MIT
- Collected
- 2026-08-04
- Default branch
- main
View the original SKILL.md
Development Workflow
Goal
Move project work from the newest accepted requirement to a verified, reviewed, reconstructable outcome with the least process that safely fits the risk.
Success Criteria
- Run the proportional phase path: Discover -> optional Decompose -> optional Plan -> optional Design -> Build -> Review -> Document. Scale phases; do not omit their applicable gates.
- Reconcile persisted state against the newest user request and current repository evidence before using its next action.
- If resume reconciliation classifies persisted state as stale, superseded, or completed,
update the framework-owned
{agent_state_dir}/task.mdbefore acting or returning; record the classification and reason, current task identity, and repaired exact next action. - Ask only about material, undiscoverable implementation choices with no safe default. Otherwise state the default and proceed.
- For medium+, create stable requirement ids with binary criteria and verification methods; small work keeps compact acceptance unless promoted by ambiguity, risk, or multiple material requirements.
- Select
plan_mode: trivial safe work usesnone, bounded small work may useinline, and medium+, risky, destructive, or scope-shaping work usesapproval_requiredand waits before Build. - Behavior changes use valid RED -> minimal GREEN -> refactor-safe verification, or record an approved non-behavior exception.
- Ordinary medium work uses one bounded edit/test executor plus independent Code Reviewer. Separate Writer and Builder/Tester only when risk or broad, noisy, environment-heavy, or explicitly independent verification requires it.
- Complete applicable Spec, quality, security, and QA routes. QA is required only by its explicit acceptance triggers.
- Build repair owns implementation/verification failures; Review owns review-fix
and fresh re-review. Document is the sole owner of
final_handoff. - Medium+ output follows
references/final-handoff.md; every accepted requirement has passed evidence or an approved exclusion.
Constraints
- Explicit user or repository artifact schemas override workflow-internal shapes; preserve exact paths, keys, types, ids, and supplied literals.
- Preserve user changes, secrets, permissions, and company data boundaries.
- Prefer repository-native commands and patterns. Do not install or call external services without approval.
- Keep scope changes explicit. Replan and seek approval when files, behavior, risk, verification, or acceptance changes materially.
- Treat SOLID, KISS, DRY, YAGNI, and patterns as evidence lenses. Refactor only for a concrete correctness, security, ownership, testing, extension, or maintainability risk.
- Exact phase markers are required only by strict mode, project policy, or user request; otherwise use concise progress updates.
Contract Loading
Read contracts/index.yaml first. Canonical contracts remain authoritative;
load only the active selector:
entryfor task inputs and triage;current_phasefor the active transition and invariants;selected_handoffbefore dispatch and return validation;completionbefore final exit.delegationbefore dispatch for indexed role/trigger fields.
Missing or invalid selectors use load_full_authoritative_file and record the
recovery. Load references/workflow-controller.md for cross-phase routing and
references/phases.md for phase mechanics. Load other references only when
their named condition applies.
Migration: workflow v4 uses subagent_trigger_scope; verification_command
remains non-empty argv string[]. assistant-review v3 owns Reviewer/QA packets
and returns final_summary / qa_evaluation_result.
Execution
- Triage size, risk, gates, state, lane, review/QA, and delegation triggers.
- Discover the relevant code and constraints; reconcile state. Create the Requirement Acceptance Map when its medium+/promotion condition applies.
- Apply
plan_mode: skip Plan fornone, keepinlineconcise, and forapproval_requireddecompose as needed and obtain approval before edits. - Build tests-first where behavior changes, verify each slice, and record deviations instead of improvising scope.
- Run Spec Review then independent quality review. Audit normally uses one
pass; review-fix normally uses one fresh post-fix re-review. Round 3+ needs
new evidence and an
additional_round_reason; round 10 is terminal. - Complete documentation and the evidence-bounded final handoff.
Output
For medium+, return status, changed areas, requirement evidence, verification, review claim, architecture decisions, manual scenarios or explicit N/A, limitations, rollback/recovery, and one next step. Small work returns status, changed areas, compact acceptance/verification, material risks, and next step.
Stop Rules
- Stop for a material unknown with no safe discoverable default.
- Stop before unapproved medium+ Build or material plan deviation.
- Stop before completion when accepted requirement evidence, validation, required review, or final-handoff fields are missing.
- Report blockers and residual uncertainty; never replace evidence with “should work,” categorical cleanliness, or guessed completion.
Alternatives
Compare before choosing
laimis91/assistant-framework
assistant-workflow
Run proportional development phases and resume persisted task state. Use to plan, build, implement, fix, migrate, refactor, or continue project artifacts.
alirezarezvani/claude-skills
app-store-optimization
App Store Optimization (ASO) toolkit for researching keywords, analyzing competitor rankings, generating metadata suggestions, and improving app visibility on Apple App Store and Google Play Store. Use when the user asks about ASO, app store rankings, app metadata, app titles and descriptions, app store listings, app visibility, or mobile app marketing on iOS or Android. Supports keyword research and scoring, competitor keyword analysis, metadata optimization, A/B test planning, launch checklist
wanshuiyin/Auto-claude-code-research-in-sleep
citation-audit
Use it for operations and research tasks; the detail page covers purpose, installation, and practical steps.
dotnet/skills
migrate-vstest-to-mtp
Migrates .NET test projects from VSTest to Microsoft.Testing.Platform (MTP). Use when user asks to "migrate to MTP", "switch from VSTest", "enable Microsoft.Testing.Platform", "use MTP runner", set OutputType=Exe only for test projects in Directory.Build.props, or mentions EnableMSTestRunner, EnableNUnitRunner, or UseMicrosoftTestingPlatformRunner. USE FOR: MTP behavioral differences vs VSTest (exit code 8, zero tests discovered, --ignore-exit-code, TESTINGPLATFORM_EXITCODE_IGNORE); centralizing