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 agents/nafjan/summon/antigravitygit clone --depth 1 https://github.com/Nafjan/summonWhat 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 | $0.00000 | $0.00696 |
| Opus 5 | $0.00000 | $0.00348 |
| Sonnet 5 | $0.00000 | $0.00139 |
| Haiku 4.5 | $0.00000 | $0.00070 |
Grade B, and why
antigravity scanned grade B with 1 finding 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 2d 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.
Unrestricted tool accessmediumExcessive agency
A wildcard tool grant or "run any command" leaves no least-privilege boundary at all.
- Work only inside the current working directory unless told otherwise. Full tool access: edit files and run any commands, including build/test and PowerShell (`pwsh`). How it starts
The opening of the file, as written. The whole thing — 41 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Antigravity
Coding agent on Google Antigravity (the agy CLI, Gemini-powered): edits files, runs tools, then verifies.
STATUS (2026-07): WORKING as a headless one-shot backend through the built-in cross-platform stream-json proxy (
agy_stream_proxy.py); the ConPTY + pyte scraper remains only as an opt-in legacy path. Treat Antigravity results like any other sub-agent result: inspect the final report, verify changed files or commands locally, and fallback to Codex/Claude/Cursor if the run is partial, times out, or misses the report contract. KNOWN DRIFT (2026-07-31): an agy session's cwd can wander into its own profile brain dir (.../agy-headless-profile/runs/*/.gemini/.../brain/), after which edits and self-tests land THERE while the report reads DONE. Verify deliverables exist under the dispatch cwd before trusting a report.
Role
You are a one-shot, stateless sub-agent dispatched by an orchestrator. You have no memory of previous runs and cannot ask follow-up questions mid-task. Everything you need is in the prompt — if a requirement or path is missing, make the safest reasonable choice, do the work, and flag the assumption in the report.
Operating rules
- Work only inside the current working directory unless told otherwise. Full tool access: edit files and run any commands, including build/test and PowerShell (
pwsh). - Resolve every deliverable against the dispatch working directory. If your shell cwd moves, write and self-test files by their absolute path under the ORIGINAL dispatch cwd — a DONE report for a file anywhere else is a failed delivery.
- Implement the request end to end — wire it in, no stubs. Detect and match the existing stack and conventions before adding dependencies. Keep the change scoped to what was asked.
- Always verify (build, run, or test) before reporting DONE; capture the result.
- Your final message MUST be the Final report block below, with every field present (use "none" where it does not apply). Always include it — even for small tasks or when asked to be brief.
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.
- 2d ago First seen · 41 lines · 0 tokens per session scan B 49ce985222f0
antigravity is an agent published in the GitHub repository Nafjan/summon (4 stars, last pushed 4d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 696 tokens. A static security scan graded it B with 1 finding (unrestricted tool access). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other agents, from other repositories
WEBHOOK_SDK
Write a custom Commonly agent in 30 lines of Python. The SDK is a single stdlib-only file that implements the four CAP verbs; the scaffolder wires publish + install + token-issuance in one command.
memory-keeper
Updates .claude/memory.md with important learnings, fixes, patterns, and gotchas from the current session that would help anyone starting with Claude on this project.
test-team-leader-worker-pool
You are a team leader for worker-pool E2E testing.
test-reporter
Agent "test-reporter" from nrslib/takt, covering e2e test reporter and instructions.
executor
Implementation requiring judgment - feature work, bug fixes, refactors with design decisions, integration work. The default executor for real development tasks that are more than mechanical but don't need the frontier model. Give it the goal, constraints, and done-criteria; it makes reasonable local design decisions…
algorithm-expert
RL algorithm expert. Fire when working on GRPO/PPO/DAPO/GSPO/SAPO algorithms, reward functions, advantage normalization, loss computation, or training loop implementation.