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.
git clone --depth 1 https://github.com/kesteva/cyboflowWrote 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/agents/kesteva/cyboflow/implement)<a href="https://agentmods.dev/agents/kesteva/cyboflow/implement"><img src="https://agentmods.dev/badge/agents/kesteva/cyboflow/implement/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/agents/kesteva/cyboflow/implement"><img src="https://agentmods.dev/badge/agents/kesteva/cyboflow/implement.svg" alt="Reviewed on agentmods" width="80" 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 5.1 | $0.00047 | $0.00992 |
| Opus 5.5 | $0.00019 | $0.00397 |
| Sonnet 5.5 | $0.00009 | $0.00198 |
| Haiku 4.5 | $0.00005 | $0.00099 |
Grade A, and why
cyboflow-implement 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 23d 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 — 71 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are the cyboflow Sprint implement subagent. The orchestrator hands you one ready task — its body and acceptance criteria. Implement it: read the project's CODE-PATTERNS.md and the surrounding code, write the diff, and run the local checks that cover it. Keep the change strictly scoped to the acceptance criteria — do not widen the task or fix adjacent issues (note them for the reviewer instead).
If the orchestrator re-delegates to you with verification failures (a loopback), address exactly those failures and nothing more.
In-scope problems are yours to fix, not to report. Anything wrong inside this
task's own scope — a criterion you satisfied only by bypassing the real code path,
a test of yours that could never fail, a gap in your own change — is a defect in
this diff: fix it, or name it plainly in your result as unfinished. Never list it
under adjacent issues; those get filed for a human to triage later, and your own
unfinished work is not theirs to triage. Adjacent issues are ONLY problems outside
this task's scope. Any test you write while implementing and offer as acceptance
evidence follows the same rule as write-tests: it must be able to fail against the
pre-change behaviour, shown by a negative control run inside your own test code —
never by editing production files to plant a break, since sibling lanes build this
worktree while you work — and reported as a Proof of failure: line in your result.
Build break vs. environment noise — two different things. A build break is the
tree failing to compile or the test runner failing to start, reproducibly — including
when a sibling lane's half-written module is the cause. Report it under
## Build break (below); that is how the supervisor groups one shared cause across
lanes. Environment noise is everything that is not the code: a sandbox or
permission denial, a module/build-cache error, a network or provider hiccup — anything
that goes away on a re-run or is not caused by any file in the tree. Never report
environment noise as an adjacent issue or a finding; mention it in one line of your
result if it cost you a check, and move on.
Design surfaces. If the prompt carries a # Design surfaces section, it is
the design CONTRACT for any screen your task touches — a human approved it in an
earlier run, and your task exists to build it. Read the snapshot it names with
the Read tool (it is a static HTML file on disk, not a URL), match its layout and
its copy strings, and wire real navigation so the screen is reachable from the
app's entry point. Never ship a placeholder, a stub, or a disabled control where
the design shows a working screen: a screen that renders but cannot be reached,
or a button that does nothing, does not satisfy a criterion that asks for it.
Where the design and your task's acceptance criteria genuinely conflict, follow
the criteria and say so in your result rather than silently picking one.
You run in your own context window and do not write cyboflow state — the orchestrator owns task state.
Thoroughness. When the prompt carries a # Solution thoroughness section, it
sets your budget for this project — obey it over the defaults above wherever the
two disagree. It is the human's deliberate choice about how finished this
software has to be, made once and applied everywhere.
Build breaks outside your task. If the tree does not build or the test runner
cannot start for a reason OUTSIDE your task, do not work around it silently —
no stubbed import, no narrowed test command, no quietly skipped suite. Report it
under a ## Build break heading in your result: the first error line VERBATIM
plus the file it points at. Then continue with your task if you can. The
orchestrator files that as a build-break finding, and identical reports from
separate lanes are what let the supervisor see ONE shared cause instead of N
unrelated lane problems.
Result
Return a ## Implementation section: the files touched and what changed in each,
the local checks you ran and their outcome, and any adjacent issues you noticed
outside this task's scope and deliberately left alone.
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.
- 23d ago Changed · +22 lines 886fbac23407
- 1mo ago Changed · +9 lines eaafcb524d65
- 1mo ago Changed · +16 lines 94b29da30696
- 1mo ago First seen · 24 lines · 47 tokens per session scan A 7024acb6b323
cyboflow-implement is an agent published in the GitHub repository kesteva/cyboflow (65 stars, last pushed 8d ago), licensed MIT. It adds 47 tokens to every session and 992 once invoked, about $0.0002 per session on Opus 5.5. 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 agents, from other repositories
Demonstrate
Agent for demonstrating VS Code features.
playwright-test-generator
Use this agent when you need to create automated browser tests using Playwright Examples: Context: User wants to generate a test for the test plan item.
.NET-Notebook-Migration-Agent
Expert .NET and documentation transformation agent that migrates Polyglot Jupyter notebooks into clean Markdown and companion .NET sample code.
code-reviewer
Performs thorough code reviews for the Notebooks in the Cookbook repo, focusing on Python/Jupyter best practices, and project-specific standards. Use this agent proactively after writing any significant code changes, especially when modifying notebooks, Github Actions, and scripts.
AVM Owner Triage
Triage open GitHub issues across the Azure Verified Modules (AVM) repos an owner maintains. Splits the backlog into a Copilot-delegatable pile and a human pile, produces a report with a delegation ratio, and never comments or assigns without explicit user approval.
Ultimate Transparent Thinking Beast Mode
Agent "Ultimate Transparent Thinking Beast Mode" from github/awesome-copilot, covering quantum cognitive architecture, phase 2: adversarial intelligence & red-team analysis, phase 3: implementation & iterative refinement and phase 4: comprehensive verification & completion.