Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/viktorbezdek/skillstack/swarm-synthesisnpx skills add viktorbezdek/skillstack --skill swarm-synthesisgit clone --depth 1 https://github.com/viktorbezdek/skillstackWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/viktorbezdek/skillstack/swarm-synthesis)<a href="https://agentmods.dev/skills/viktorbezdek/skillstack/swarm-synthesis"><img src="https://agentmods.dev/badge/skills/viktorbezdek/skillstack/swarm-synthesis.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00143 | $0.02486 |
| Opus 5 | $0.00072 | $0.01243 |
| Sonnet 5 | $0.00029 | $0.00497 |
| Haiku 4.5 | $0.00014 | $0.00249 |
Grade A, and why
swarm-synthesis scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 257 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Swarm Synthesis
The swarm's value is in the dissent. Synthesis that hides disagreement to "give a clear answer" loses what the multi-perspective brainstorm bought you.
This skill turns a pile of persona contributions into an actionable artifact. The artifact preserves dissent explicitly, surfaces consensus where it exists, names the open questions, and proposes a synthesized next move — without flattening the perspectives.
Core principle
Synthesis is curation, not consensus. The synthesizer's job is to make the swarm's structure visible to the user — not to decide who's right. Disagreement is data; preserve it.
Secondary principle: end with a recommended next move. A synthesis that ends in "well, perspectives differ" puts the burden on the user. A synthesis that ends with "given the disagreement, the next move is to gather X evidence" is actionable.
The synthesis artifact
Standard format:
## Brainstorm: [topic]
[1-2 sentence framing — what was being brainstormed]
### Personas in the swarm
[List of 4-8 personas spawned + 1-line rationale per inclusion]
### Consensus
- [Point all personas (or near-all) agreed on, with attribution to who raised it first]
- [Another consensus point]
[3-5 items, ordered by importance]
### Dissent
- **[Topic of disagreement]**: PM said [position]; Skeptic said [opposing position]; the underlying disagreement is [the real question, not the surface position]
- **[Topic of disagreement]**: ...
[2-4 items, each naming the substantive disagreement, not just the surface clash]
### Open questions (nobody could answer)
- [Question that surfaced but no persona could resolve]
- [Question]
[1-3 items, ideally with "what evidence would resolve this"]
### Recommended next move
- [Synthesized direction with reasoning]
- [Alternative if the recommended move is rejected]
### Per-persona contributions (collapsed)
<details>
<summary>PM perspective</summary>
[Full PM output]
</details>
<details>
<summary>Engineer perspective</summary>
[Full Engineer output]
</details>
[etc.]
What ships with it
5 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 3d ago First seen · 257 lines · 143 tokens per session scan A d77b68cbdfed
swarm-synthesis is a skill published in the GitHub repository viktorbezdek/skillstack (11 stars, last pushed 3mo ago), licensed MIT. It adds 143 tokens to every session and 2,486 once invoked, about $0.0007 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
gossipcat
Use when installing or setting up gossipcat multi-agent orchestration (parallel review, consensus, adaptive dispatch) in Claude Code or Cursor. Installs the gossipcat MCP server and hands off to gossipstatus() for all live rules.
Elephant Agent
Complete guide to what Elephant Agent is, how to use it from the CLI, how herd, wake, Personal Model, proactive curiosity, skills, and tools fit together, and how to explain Elephant Agent clearly to users, contributors, or evaluators.
pytorch-test-refactoring
Orchestrate PyTorch test file refactoring to decouple tests from specific hardware accelerators. Use this skill whenever the user asks to refactor, decouple, or reorganize a PyTorch test file for cross-accelerator compatibility, or when they ask to apply the test decoupling workflow to a specific test file. Triggers…
refactor-test-decoupling
Refactor PyTorch test files to decouple tests from specific hardware accelerators using three strategies: CPU-only (standalone classes with instantiateparametrizedtests), device-agnostic (device-parametrized classes with instantiatedevicetypetests), and device-specific (single-accelerator standalone classes). Pay…
review-test-refactoring
Review PyTorch test refactoring for correctness and completeness against the decoupling standards defined in the refactor-test-decoupling skill. Accepts a test file path (whole-file review), a PR URL, a git diff, or a branch name. Use this when asked to review a test refactoring PR, check a test decoupling change…
adversarial-review
Runs an adversarial review loop where a challenger agent critically reviews an implementer's output, driving iterative improvement until the challenger issues a PASS verdict. Never triggered automatically; must be explicitly called by the user (2 persistent agents, up to 10 rounds).