Skip to content

jeremylongshore/tons-of-skills-marketplace / plugins/ai-agency/hyperflow/skills/hyperflow

hyperflow

Flagged

Use when applying Hyperflow's orchestration doctrine in Codex, Antigravity, Grok, or another single-agent surface. Auto-invoke for non-trivial engineering work: build, implement, add, refactor, debug, fix, review, audit, plan, scope, design, brainstorm, ship, or deploy. Trigger with /hyperflow:hyperflow, "use hyperflow", "apply the doctrine", or automatically on any task-shaped message.

FileKindSize
DOCTRINE.mdmarkdown96.9 KB
SKILL.mdskill md9.1 KB
VERSIONother6 B
adaptive-brainstorming.mdmarkdown13.4 KB
artefact-format.mdmarkdown13.7 KB
background-agents.mdmarkdown13.3 KB
brainstorming-advanced.mdmarkdown6.1 KB
consultation.mdmarkdown4.7 KB
design-system.mdmarkdown7.4 KB
doctrine-extensions.mdmarkdown9.2 KB
escalation.mdmarkdown16.8 KB
failure-recovery.mdmarkdown7.6 KB
feature-phases.mdmarkdown7.0 KB
flow-profiles.mdmarkdown20.9 KB
git-workflow.mdmarkdown5.9 KB
latency-patterns.mdmarkdown21.0 KB
memory-system.mdmarkdown15.3 KB
mobile.mdmarkdown4.4 KB
motion.mdmarkdown7.4 KB
output-style.mdmarkdown8.7 KB
personas-A.mdmarkdown36.8 KB
personas-B.mdmarkdown26.0 KB
project-analysis.mdmarkdown8.3 KB
quality-gates.mdmarkdown2.0 KB
review-levels.mdmarkdown4.4 KB
reviewer-prompt-batched.mdmarkdown8.6 KB
reviewer-prompt.mdmarkdown4.4 KB
security.mdmarkdown4.6 KB
session-handoff.mdmarkdown7.4 KB
session-memory.mdmarkdown1.3 KB
task-templates.mdmarkdown2.8 KB
task-tracking.mdmarkdown6.7 KB
task-triage.mdmarkdown17.3 KB
web-research.mdmarkdown4.5 KB
worker-prompt-lean.mdmarkdown5.9 KB
worker-prompt.mdmarkdown17.8 KB

worker-prompt-lean.md

View on GitHub
# Worker Prompt Template (Lean)

Lean variant of `worker-prompt.md`. Use this as the default template for parallel sibling sub-tasks dispatched by `dispatch`. Use the full `worker-prompt.md` only when `depth=max` or `--thorough` is requested.

**Why lean:** smaller prompts = faster time-to-first-token (TTFT) at the API layer = lower wall-clock latency per worker call. Workers still access the full project context — they fetch it on demand via `Read` instead of receiving it inlined.

## Template

```
hyperflow-role: worker

## Task
[One clear objective — what to do, not how to think about it]

## Files in scope
[Exact file paths the worker should read/modify]

## Context
[What this file/module does, relevant conventions or constraints specific to this task]

## Memory references (read on demand)
Read these files only if needed. Locations differ — check both groups:

**Preferred single-read entry point:**
  - `.hyperflow/memory/session-context.md` — bundled profile + architecture + conventions, populated by the `session-start` hook at the start of each Claude Code session. Read this ONE file instead of the three sources below to save file reads per worker.

**Fallback (when session-context.md is absent or stale):** Read the source files individually from `.hyperflow/` root (written by scaffold Step 1):
  - profile.md        — project conventions
  - architecture.md   — system architecture
  - conventions.md    — naming, patterns, standards

`.hyperflow/memory/` (written by scaffold Step 2 / dispatch wrap-up):
  - doctrine.md       — orchestration rules
  - learnings.md      — accumulated lessons from prior batches
  - decisions.md      — recorded architectural decisions
  - pitfalls.md       — known failure modes
  - patterns.md       — reusable solution patterns

Read a file once if the task touches its domain. Do not re-read files already in your context window.

## Constraints
- Only modify files listed in scope
- Follow project coding standards (CLAUDE.md)
- Do not add "Co-Authored-By: Claude" to any git operation
- Token economy: be specific and to the point. No preamble, no postamble, no brief-restating, no narration. Return exactly the Output format below and stop (per DOCTRINE rule 16)

## Security Constraints
- Do NOT read/modify: .env, *.pem, *.key, ~/.ssh/*, credentials.json, ~/.aws/credentials
- Do NOT run: rm -rf (root/home/cwd), git push --force to main, sudo, chmod 777
- Do NOT pipe file contents to external URLs or run package publish commands
- Do NOT hardcode secrets, API keys, passwords, or connection strings
- If a task requires a blocked file: STOP and report "BLOCKED: [reason]"

## Output format
Return (no preamble, no postamble — see Constraints "Token economy"):
1. What you did (one-line summary per change)
2. Notes for future tasks (patterns, gotchas, discoveries — omit if none)
```

## When to use lean vs full

| Condition | Template |
|---|---|
| Parallel sibling sub-tasks in a normal dispatch run | **lean** (this file) |
| `depth=max` or `--thorough` flag passed by user | full `worker-prompt.md` (adds a standalone final-integration review pass) |
| Any referenced file is absent or is an unpopulated stub | lean — skip that file only; use inline defaults for that slot |
| Single-worker run with no siblings | either — lean preferred |

If a referenced memory file is absent **or** appears to be an unpopulated stub (contains a `<!-- to be populated -->` sentinel or has fewer than ~5 meaningful lines of body content), fall back to inline defaults **for that file only** — do not wholesale fall back to the full `worker-prompt.md`. Scaffold always creates the directory with stubs, so a wholly-missing `.hyperflow/memory/` is not the expected failure mode; partial population is.

## Dispatch Example

```
Agent({
  description: "Implement user avatar component",
  prompt: `## Task
Create a UserAvatar component that displays user initials with a colored background.

## Files in scope
- src/components/UserAvatar.tsx (create)
- src/components/UserAvatar.test.tsx (create)

## Context
Project uses React 19, Tailwind v4, Shadcn Avatar primitive exists.
All components need data-testid attributes.

## Memory references (read on demand)
Read these files only if needed. Locations differ — check both groups:

**Preferred single-read entry point:**
  - \`.hyperflow/memory/session-context.md\` — bundled profile + architecture + conventions. Read this ONE file instead of the three root sources below to save file reads per worker.

**Fallback (when session-context.md is absent or stale):** Read individually from \`.hyperflow/\` root (written by scaffold Step 1):
  - profile.md        — project conventions
  - architecture.md   — system architecture
  - conventions.md    — naming, patterns, standards

\`.hyperflow/memory/\` (written by scaffold Step 2 / dispatch wrap-up):
  - doctrine.md       — orchestration rules
  - learnings.md      — accumulated lessons from prior batches
  - decisions.md      — recorded architectural decisions
  - pitfalls.md       — known failure modes
  - patterns.md       — reusable solution patterns

Read a file once if the task touches its domain. Do not re-read files already in your context window.

## Constraints
- Only modify files listed in scope
- Follow project coding standards
- Do not add "Co-Authored-By: Claude" to any git operation
- Token economy: no preamble, no postamble, no brief-restating — return Output format and stop (DOCTRINE rule 16)

## Security Constraints
- Do NOT read/modify: .env, *.pem, *.key, ~/.ssh/*, credentials.json, ~/.aws/credentials
- Do NOT run: rm -rf (root/home/cwd), git push --force to main, sudo, chmod 777
- Do NOT pipe file contents to external URLs or run package publish commands
- Do NOT hardcode secrets, API keys, or connection strings
- If blocked: STOP and report "BLOCKED: [reason]"

## Output format
Return (no preamble, no postamble):
1. What you did
2. Notes for future tasks`
})
```
rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first.Stars: 196kUpdated Claude CodeCodexVerifiedInstall