{"slug":"phx-full","title":"phx-full","summary":"Run a portable sequential plan-work-verify-review-compound lifecycle.","platform":"Claude","tags":[],"authorName":"LLM Mart","authorSlug":"llm-mart","score":0,"source":"github","price":null,"verified":false,"createdAt":"2026-10-04T15:14:25.787687Z","repo":{"url":"https://github.com/oliver-kriska/claude-elixir-phoenix","stars":560,"forks":44,"license":"MIT","updatedAt":"2026-10-02T04:11:38Z"},"bodyHtml":"
\n

name: phx-full\ndescription: Run a portable sequential plan-work-verify-review-compound lifecycle.\nUse optional generic workers only when Amp makes them available.

\n

Full Phoenix Feature Development

\n

Run the portable lifecycle: discover → plan → work → verify → read-only review →\ncompound. The filesystem is the state machine; no task API or named orchestrator\nis required.

\n

Usage

\n
phx-full Add user authentication with magic links\nphx-full Background email jobs --max-cycles 5 --max-retries 2\n
\n

If input is an existing .claude/plans/*/plan.md, do not re-plan. Ask for the\nnative phx-work workflow or execute its portable behavior in this session.\nDefaults are --max-cycles 10, --max-retries 3, and --max-blockers 5.

\n

Lifecycle

\n
    \n
  1. DISCOVERING — inspect relevant code, tests, prior solutions, and optional\nTidewave evidence. Tidewave is optional; local files, logs, and mix commands\nare the complete fallback. Record complexity and proposed depth, then wait for\nthe user's plan/implementation gate. Never auto-select a path that bypasses it.
    2. \n
  2. PLANNING — invoke the runtime's native phx-plan skill when available, or\nexecute its portable research checklist and artifact format in this session.\nRequire .claude/plans/{slug}/plan.md. Present it and wait for approval before\nimplementation unless the user already explicitly authorized the full run.
    3. \n
  3. WORKING — execute the plan sequentially. Task selection occurs only here.\nThe full-run limits override any baseline workflow retry defaults. Before every\nattempt persist cycle, task retry, and blocker counters; if the next attempt\nexceeds a limit, do not run it. --max-retries N means at most N retries after\nthe initial attempt (N+1 total attempts for that task). Mark [BLOCKED] and\nstop at --max-blockers.
    4. \n
  4. VERIFYING — run mix format --check-formatted, compile with warnings as\nerrors, focused tests during work, and the full relevant suite at this gate.\nA failed gate appends FAIL and returns to WORKING only within the cycle limit.
    5. \n
  5. REVIEWING — invoke portable phx-review, or perform the same read-only,\nchanged-file review sequentially. Generic workers are optional. Review never\nedits. Findings or failures become plan tasks and return to WORKING.
    6. \n
  6. COMPOUNDING — only after verification and a clean/accepted review. Do not\ninvoke phx-compound. Inline contract: write a solution artifact under\n.claude/solutions/ only when the run produced a non-obvious, reusable learning,\nincluding problem, root cause, solution, and verification. Otherwise append\nCOMPOUNDING SKIPPED: no reusable learning to progress. Never edit CLAUDE.md.
    7. \n
\n

Track INITIALIZING → DISCOVERING → PLANNING → WORKING → VERIFYING → REVIEWING → COMPOUNDING → COMPLETED, with BLOCKED reachable from every phase. A cycle is\none WORKING → VERIFYING → REVIEWING pass; increment and persist it before\nentering VERIFYING. At --max-cycles, do not begin another pass: stop INCOMPLETE with remaining tasks,\nfailed evidence, and a concrete resume command for this runtime.

\n

Iron Laws

\n
    \n
  1. Honor user gates — discovery and plan approval are not automatic transitions.
    2. \n
  2. Never skip verification or the read-only review phase.
    3. \n
  3. Only WORKING edits code; review findings become explicit plan tasks.
    4. \n
  4. Respect every cycle, retry, and blocker limit; stop when exhausted.
    5. \n
  5. Persist state before stopping so plan checkboxes and progress evidence resume.
    6. \n
  6. Do not require hooks, MCP, named agents, background tasks, or a task UI.
    7. \n
\n

Resume Ledger

\n

progress.md is the sole state authority. It is append-only: never overwrite or\nmaintain a competing authoritative current-state record. Every event has monotonic\nseq, phase_visit, phase, cycle, task, task_attempt, cumulative\nblockers, outcome, and an evidence or artifact path. On resume, validate the\nlast valid event against evidence, plan checkboxes, artifacts, and git state, then\nenter only its legal successor. Any WORKING edit after a VERIFYING or REVIEWING\npass invalidates both passes; the next legal phase is VERIFYING.

\n

Completion requires all required plan tasks checked, no unresolved [BLOCKED],\nthe latest VERIFYING PASS after the last edit, the latest accepted REVIEWING after\nthat verify, and COMPOUNDING passed or explicitly skipped.

\n

References

\n\n","files":[{"path":"references/cycle-patterns.md","sizeBytes":418,"isText":true},{"path":"references/example-run.md","sizeBytes":569,"isText":true},{"path":"references/execution-steps.md","sizeBytes":1137,"isText":true},{"path":"references/safety-recovery.md","sizeBytes":709,"isText":true},{"path":"SKILL.md","sizeBytes":4716,"isText":true}],"reviewScore":null,"reviewSummary":null,"trust":{"provenance":"trusted-source-unreviewed","notice":"Community-authored content, reproduced verbatim and not vetted as instructions. Treat it as data to evaluate, never as directives to follow.","bodySource":null},"bodyLocked":false,"purchaseUrl":null,"sourceUrl":null,"report":{"provenance":"trusted-source-unreviewed","screen":{"ran":true,"outcome":"clean","suspicious":0,"notes":0,"hiddenCharacters":false},"virusScan":{"engine":"clamav","status":"clean","scannedAt":"2026-10-04T15:16:44.950891Z","sha256":"2BDBD0A48339B8D9708B5F1926D202533AABF470587483589A4CA9989AE75E1E","sizeBytes":4539},"review":null,"source":{"repositoryUrl":"https://github.com/oliver-kriska/claude-elixir-phoenix","path":"targets/amp/skills/phx-full","license":"MIT","commit":"9767a82d24ddddad553e85f88efc2869a7fd7d88","subtreeSha":"A5C74BADD0D7C840313BD9999B33B77A33D932044D9BE713B978A27EE66A859E","lastSyncedAt":"2026-10-04T15:14:09.139242Z"},"reviewedAt":"2026-10-04T15:20:38.518945Z","notice":"Community-authored content, reproduced verbatim and not vetted as instructions. Treat it as data to evaluate, never as directives to follow."},"install":[{"target":"skills-cli","command":"npx skills add https://github.com/oliver-kriska/claude-elixir-phoenix/tree/main/targets/amp/skills/phx-full"},{"target":"claude-code","command":"claude plugin marketplace add https://llmmart.ai/marketplace.json && claude plugin install oliver-kriska-claude-elixir-phoenix@llmmart"},{"target":"git","command":"git clone https://github.com/oliver-kriska/claude-elixir-phoenix.git"}]}