abofles

Plugin abofles from marketplace sajadh92 in Sajadh92/claude-plugins. Commands read from the repo's manifest — 9 days ago.

Skills (1)

• abofles

  Abo Fles (ابو فليس) — picks the model (haiku / sonnet / opus / fable) and effort (low..max) for every subagent in coding work from a topic × side matrix, and suggests the main-session model. TRIGGER once per session, before the first coding task or Agent tool call (it stays in context after that — n

Slash commands (10)

• /abofles:agents

  Agent gate — show or set how many agents a task may spawn: auto (sized by tier), N (this task), off, or ask again

• /abofles:bill

  Show what this session's subagents cost — the unit index per model and side, and the real tokens and dollars — and what Abo Fles saved versus running them all on one model

• /abofles:guide

  What to do next — one prioritised line from the ledger, the memo and the context size (open FAIL, unasked question, owner decisions, escalation, background agents, memo/compact, budget)

• /abofles:init

  Add an Abo Fles override block to this project's CLAUDE.md, with tier suggestions based on what the code does

• /abofles:memo

  Update the Now section of the work memo (task, done, left, decisions, next, verify) so the session can be compacted or paused without losing the thread; then run /compact yourself

• /abofles:pane

  The Abo Fles pane (resume, spawns, overrides) — toggled by /abofles-pane (abofles-guide mod); without the mod, prints the resume block and the spawn history as text

• /abofles:resume

  Where was I? Print the resume block from the work memo (now · next · needs you · last receipts) without re-reading anything; --resolve <n> <note> ticks an item (the note is required: the evidence), --note <text> records something settled in chat

• /abofles:route

  Dry run — print the Abo Fles route lines (model, effort, reply cap) for a task without spawning anything

• /abofles:spawns

  One line per spawn of this session — time, side, status, confidence, model·effort, tokens, $ and saved — plus the totals line (the pane's Spawns card as text)

• /abofles:status

  Session summary from the ledger: spawns per model as bars, confidence, and the open follow-ups (partial, blocked, low, ATTENTION, FAIL); --resolve <id> --trace <evidence> closes one

Subagents (19)

• abofles-analyst

  Abo Fles Analyse & Find Gaps agent (side Analyse). Judges code, finds bugs, risks and gaps, compares options, suggests a better approach. Returns ranked findings with a fix per finding. Use for \"review X for risks\", \"find the bug in Y\", \"compare approaches\", \"what is missing in Z\".

• abofles-cleancoder

  Abo Fles Clean Code reviewer (side Verify, tag Clean). Reviews a finished change for readability and structure: naming, duplication, dead code, function size, module boundaries — without proposing behaviour changes. Returns PASS or FAIL with evidence. Use after Verify on 🟠/🔴 changes in Code Health

• abofles-compliance

  Abo Fles Compliance reviewer (side Verify, tag Compliance). Reviews a finished change for PII exposure, consent, retention and deletion, audit trail completeness, and legal or regulatory text. Returns PASS or FAIL with evidence. Use after Verify on changes in Data Privacy, Audit Logging and any topi

• abofles-data

  Abo Fles Data reviewer (side Verify, tag Data). Reviews a finished data-pipeline, analytics, report or dashboard change for correctness: idempotent loads and backfills, metric definitions, aggregation and join correctness, double counting, freshness and data contracts. Returns PASS or FAIL with evid

• abofles-dba

  Abo Fles DBA reviewer (side Verify, tag DBA). Reviews a finished database change — migrations, schema, queries, indexes, transactions, replicas, backups — for safety, reversibility, locking and performance under production data. Returns PASS or FAIL with evidence. Use after Verify on 🟠/🔴 changes i

• abofles-explainer

  Abo Fles Explainer agent (side Explain). Takes a Judge's HANDOFF (question, evidence, why it matters) on a T3–T4 decision and writes the question the user will actually understand: plain words, options with consequences, a recommendation, and a table, example or diagram when needed.

• abofles-finance

  Abo Fles Finance reviewer (side Verify, tag Finance). Reviews a finished change for money correctness: rounding and precision, currency handling, ledger invariants, idempotency and double entry, limits and fees. Returns PASS or FAIL with evidence. Use after Verify on 🟠/🔴 changes in Money & Finance

• abofles-git

  Abo Fles Git agent (side Work, tag Git). Owns every git and GitHub action: inspects the repo first, then branches, worktrees, commits, merges, rebases, conflict resolution, pull requests, tags and releases, with claims so several sessions and agents can share one project. Returns the actions run, th

• abofles-judge

  Abo Fles Judge agent (side Judge). Runs before any question reaches the user: answers it from the code, comments, tests, docs, ticket text given in the prompt or the memo with evidence — or, when it is really the owner's call, explains it in plain words (T1–T2) or hands the evidence to the Explainer

• abofles-ops

  Abo Fles Ops reviewer (side Verify, tag Ops). Reviews a finished infrastructure, pipeline, deployment, server or network change for blast radius, rollback path, secrets exposure, drift and environment parity. Returns PASS or FAIL with evidence. Use after Verify on 🟠/🔴 changes in DevOps, Cloud, Edg

• abofles-perf

  Abo Fles Perf reviewer (side Verify, tag Perf). Reviews a finished change for performance: hot paths and complexity, N+1 and round-trips, allocations and blocking calls, caching, and benchmark or load evidence before and after. Returns PASS or FAIL with evidence. Use after Verify on 🟠/🔴 changes in

• abofles-planner

  Abo Fles Plan & Design agent (side Plan). Writes an implementation plan, design, or migration strategy before code changes. Returns numbered steps with files, risks, and a verification step. Use for \"plan how to add X\", \"design the change for Y\", \"migration strategy for Z\".

• abofles-qa

  Abo Fles QA reviewer (side Verify, tag QA). Checks a finished change against the ticket's acceptance criteria as a user would: flows end to end, regressions on neighbouring features, error and empty states. Returns PASS or FAIL with evidence. Use after Verify on 🟠/🔴 changes in Business Logic, Inte

• abofles-reader

  Abo Fles Read & Collect agent (side Read). Finds files, maps code, collects facts, lists usages. Returns facts only, no opinions, no fixes. Use for \"where is X\", \"what calls Y\", \"list all uses of Z\", \"map this folder\".

• abofles-security

  Abo Fles Security reviewer (side Verify, tag Security). Reviews a finished change for OWASP top-10 classes, broken authorization and IDOR, secrets, injection, crypto misuse and session handling. Returns PASS or FAIL with evidence. Use after Verify on 🟠/🔴 changes in Security, Money, DevOps, Data an

• abofles-tester

  Abo Fles Tester (side Verify, tag Tester). Runs the test suite for a finished change and writes the missing edge-case tests (zero, null, duplicate, concurrent, boundary), then reports PASS or FAIL with the run output. Use after Verify on 🟠/🔴 changes; the only reviewer that edits files, and only te

• abofles-uiux

  Abo Fles UI/UX reviewer (side Verify, tag UIUX). Reviews a finished UI change for accessibility, RTL and localisation, responsive layout, design tokens and copy — with the Better Design and UX Pilot connectors when they are connected, a checklist otherwise. Returns PASS or FAIL with evidence. Use af

• abofles-verifier

  Abo Fles Verify & Review agent (side Verify). Checks finished work against acceptance criteria: runs tests, reviews the diff, confirms docs match code. Returns PASS or FAIL with evidence. Use for \"verify the change to X\", \"review this diff against the plan\", \"confirm Y still works\".

• abofles-worker

  Abo Fles Act & Work agent (side Work). Writes or edits code, tests, docs, and configs from a plan or findings. Returns the changed-files list, one line per change, and what was run. Use for \"implement step X\", \"apply these fixes\", \"write the tests for Y\".

MCP servers (4)

• atlassian

  connects to: https://mcp.atlassian.com/v2/mcp

• context7

  connects to: https://mcp.context7.com/mcp

• better-design

  connects to: https://better-design.com/api/mcp

• ux-pilot

  connects to: https://mcp-server-485825078314.us-central1.run.app/mcp/ui

Hooks (12)

• SessionStart

  echo 'Abo Fles (ابو فليس) — plugin abofles — is active: in coding work, invoke the abofles skill once per session, before the first coding task or subagent (it stays in context; do not re-invoke it pe

• SessionStart

  bash "${CLAUDE_PLUGIN_ROOT}/scripts/py" --hook "${CLAUDE_PLUGIN_ROOT}/hooks/run.py" start

• PreToolUse

  bash "${CLAUDE_PLUGIN_ROOT}/scripts/py" --hook "${CLAUDE_PLUGIN_ROOT}/hooks/run.py" guard

• PreToolUse

  bash "${CLAUDE_PLUGIN_ROOT}/scripts/py" --hook "${CLAUDE_PLUGIN_ROOT}/hooks/run.py" qgate

• PreToolUse

  bash "${CLAUDE_PLUGIN_ROOT}/scripts/py" --hook "${CLAUDE_PLUGIN_ROOT}/hooks/run.py" pre

• UserPromptSubmit

  bash "${CLAUDE_PLUGIN_ROOT}/scripts/py" --hook "${CLAUDE_PLUGIN_ROOT}/hooks/run.py" prompt

• Stop

  bash "${CLAUDE_PLUGIN_ROOT}/scripts/py" --hook "${CLAUDE_PLUGIN_ROOT}/hooks/run.py" receipt

• SubagentStop

  bash "${CLAUDE_PLUGIN_ROOT}/scripts/py" --hook "${CLAUDE_PLUGIN_ROOT}/hooks/run.py" stop

• PostToolUse

  bash "${CLAUDE_PLUGIN_ROOT}/scripts/py" --hook "${CLAUDE_PLUGIN_ROOT}/hooks/run.py" decision

• PostToolUse

  bash "${CLAUDE_PLUGIN_ROOT}/scripts/py" --hook "${CLAUDE_PLUGIN_ROOT}/hooks/run.py" post

• PostToolUse

  bash "${CLAUDE_PLUGIN_ROOT}/scripts/py" --hook "${CLAUDE_PLUGIN_ROOT}/hooks/run.py" sendmsg

• PostToolUse

  bash "${CLAUDE_PLUGIN_ROOT}/scripts/py" --hook "${CLAUDE_PLUGIN_ROOT}/hooks/run.py" edit

Similar plugins