Menu
SkillsMP has collected 23 skills from AleSaiani/agentflow. Open a skill to review its source and details.
Showing 23 of 23 collected skills.
Apply ONE operation to EVERY item of a list — the **map** (N→N). The operation is a prompt; items come from a JSON list, a markdown checklist, or another run's output. Runs in parallel subagents or inline in the main thread, with persistent state and…
Run a declarative workflow-file end to end: init a `/agentflow:pipe` from a JSON `WorkflowSpec`, then drive it to completion — auto-running every bash/json/deterministic stage and dispatching agents only where genuinely needed. Use `--dry-run` to preview the…
Deep code-audit recipe: discover files in a target → review each (LLM, cached) → partition by component → executive digest with hotspots and recurring patterns. A layer-3 recipe over /agentflow:pipe (the 6-stage pipeline and caching are detailed in the body).…
List the reusable workflows authored in this workspace — every `workflows/<name>/` folder, with its format (WORKFLOW.md / workflow.json), stage count, declared params, and description (read from either format). The "what workflows do I have to run" catalog.…
Workspace dashboard: lists every active framework run across all primitives, highlights blockers (stuck items, exhausted auto-continue caps), shows cumulative cost, and suggests the next action. The natural command to run after re-opening Claude Code in a…
Do a multi-step DETERMINISTIC operation ONCE, ad-hoc — without saving a workflow file. You describe the outcome; `do` designs an inline pipeline (code/bash stages where possible, an LLM step only where real judgment is needed — the determinism boundary),…
Answer "how do I … with Agent Flow?" — map a plain-language intent to the right command(s) and a copy-paste recipe, grounded in the installed docs and skills (never invented). It EXPLAINS; it does not execute. USE for questions about USING Agent Flow: "how…
Read-only inspector over runs of any primitive (enumerate / foreach / group / reduce / iterate / pipe). Lists runs, shows a run's status, draws a /agentflow:pipe child tree, aggregates budget across a run and its children, prints a timeline. Never mutates…
Observe and CONTROL the engine's in-flight jobs: list every top-level run in the order the Stop hook will advance them (the queue), pause/resume a single job, pause/resume the WHOLE engine, delete a run, set a job's priority, and garbage-collect finished runs…
Generate a list of items from a higher-level spec — the **unfold** (1→N) primitive. Turn an outline into chapters, a feature into tasks, a goal into a checklist. Output is an items.json array that `/agentflow:foreach` (map) or `/agentflow:group` (partition)…
Partition N items into K groups by key — the partition step. Output is items.json-compatible: feed it straight to /agentflow:foreach to process per group. Methods: path-prefix / regex / jsonpath (deterministic) or llm-classify (semantic, when no deterministic…
Chronological history of Agent Flow runs in this workspace — every run across all primitives, most recent first, with status, agent count, and cost. The "what have I run, and when" log. Read-only, never mutates state. USE when the user asks "what did I run",…
Directed outbox/inbox messaging between concurrent Agent Flow instances or sessions. One instance sends to a named box, another receives from it; `recv` claims the oldest message with an atomic rename, so two receivers never get the same message (FIFO, no…
Send a notification — a webhook POST (Slack/Discord-compatible) and/or a best-effort OS desktop notification — so the user is pinged when a long-running flow finishes or hits a milestone, even if they've stepped away from the terminal. Zero-dep; never fails…
Run an ordered pipeline of stages — each a bash command, a json write, or another primitive (/agentflow:enumerate, /agentflow:foreach, /agentflow:group, /agentflow:reduce, or a loop). The composer: it holds no map/fold/loop of its own; it sequences the…
A lock-free shared work queue that MANY workers (terminals/sessions) can drain at the same time without double-processing. Each item is a file claimed by an atomic rename, so concurrent claims are safe with no locks. Supports dynamic enqueue (`add`), a…
Collapse N results into 1 digest (markdown or JSON) via a single agent — the **fold** (N→1). USE when finished work needs synthesizing into one artifact — a /agentflow:foreach run, several result files, or inline data: "summarize all these", "roll this up",…
Run a stage a FIXED number of times — a bounded count loop (`for i in 1..N`). One iteration per turn; the Stop hook resumes the next across turns. For a loop with a semantic stop condition, use `/agentflow:until` (do…until) or `/agentflow:while` (while…do)…
Run ONE prompt once and capture its structured output — the single LLM/work unit between /agentflow:reduce (one agent over N inputs) and /agentflow:foreach (one agent per item). Pick where it runs with `--runtime`: `main` (inline), `subagent` (one Agent), or…
Loop a stage and stop once a predicate is satisfied — `do … until (cond)`: the body runs FIRST, then the condition is checked. The loop always runs at least once. Full loop safety: hard cap on iterations, convergence detection (stops if output stops…
Loop a stage WHILE a predicate holds, checking the condition BEFORE each run — `while (cond) { body }`. If the condition is already false up front, the body never runs. Full loop safety: hard cap, convergence detection, kill switch. USE when the guard must be…
Run a **repeatable checklist / to-do list** from a markdown file of `- [ ]` items — execute the open items, tick them `[x]`, and write the file back. Re-running resumes only the still-open items, so the file IS the durable, repeatable state. USE for a generic…
Author a reusable workflow-file by composing the Agent Flow primitives (enumerate, foreach, group, reduce, iterate) and bash/json stages into a declarative JSON `WorkflowSpec`, then validate it — ready to run with `/agentflow:run-workflow`. The builder…