Borrowing it
Nothing to install: this file belongs to EastStarAI/sanad-agent. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/EastStarAI/sanad-agent/main/.agents/skills/sanad-subagent-developer/SKILL.mdgit clone --depth 1 https://github.com/EastStarAI/sanad-agentWrote 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/eaststarai/sanad-agent/sanad-subagent-developer)<a href="https://agentmods.dev/skills/eaststarai/sanad-agent/sanad-subagent-developer"><img src="https://agentmods.dev/badge/skills/eaststarai/sanad-agent/sanad-subagent-developer/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/skills/eaststarai/sanad-agent/sanad-subagent-developer"><img src="https://agentmods.dev/badge/skills/eaststarai/sanad-agent/sanad-subagent-developer.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 2 findings, up to medium
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- medium Agent Snooping · line 13 Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.Fix: Remove all code or instructions that list or read other skills' files or directories. Skills should operate independently; cross-skill access is a privilege escalation.
- medium Agent Snooping · line 37 Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.Fix: Remove all code or instructions that list or read other skills' files or directories. Skills should operate independently; cross-skill access is a privilege escalation.
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 | $0.00027 | $0.00930 |
| Opus 5.5 | $0.00011 | $0.00372 |
| Sonnet 5.5 | $0.00005 | $0.00186 |
| Haiku 4.5 | $0.00003 | $0.00093 |
Grade A, and why
Sanad Subagent Developer 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 21d 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 — 57 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Sanad Subagent Developer Protocol
Execute one well-defined task in an isolated worktree, prove its Definition of Done, and return a reviewable result without contaminating another active runtime.
Role Boundaries
You are the scoped implementer in the three-role workflow (orchestrator / implementer / independent reviewer). You own the assigned change inside its worktree and brief: implement, verify, and report. You do not perform the final independent review, and you do not stand in for the orchestrator or run its review/CI-repair work. Merge, Ready conversion, protected labels, and destructive cleanup remain authorization boundaries unless the brief granted them up front.
1. Read the Plan and Create Isolation
-
Read the assigned task plan from
docs/plans/tasks/<task_id>-<task_name>.md. -
Load
.agents/skills/sanad-agentic-developer/SKILL.mdand follow its isolated-worktree mode. -
Create and enter the dedicated worktree before editing, testing, or launching processes:
git worktree add .agent/worktrees/<task_id>-<branch_name> <branch_name> -
Use
sanad-devfrom inside that worktree for runtime work. Do not offset ports, setSANAD_STATE_HOME, construct a relativeSANAD_HOME, or edit tracked runtime configuration manually. The launcher allocates worktree-specific daemon/VM ports, creates one worktree-scopedSANAD_HOME, isolates client preferences, and selects the matching runtime automatically. -
Use direct component commands only when diagnosing
sanad-devitself or one process in isolation.
2. Targeted Coding
- Modify only the scope established by the task plan and the nearest owning contracts.
- Read every applicable
AGENTS.mdbefore editing its directory. - Use FVM for every Flutter/Dart command, keep paths workspace-relative, and preserve established ownership boundaries.
- Update the corresponding active page under
docs/in the same task. Update anAGENTS.mdonly when a durable law, invariant, or ownership boundary changes or becomes stale. - Do not overwrite unrelated dirty work already present in the assigned worktree.
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.
- 21d ago Changed fcb7209dbdda
- 1mo ago Changed · +4 lines 71c7ab15a695
- 2mo ago First seen · 53 lines · 27 tokens per session scan A 457a0791cf92
Sanad Subagent Developer is a skill published in the GitHub repository EastStarAI/sanad-agent (47 stars, last pushed yesterday), licensed MIT. It adds 27 tokens to every session and 930 once invoked, about $0.0001 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.
Sanad Subagent Developer compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| Sanad Subagent DeveloperEastStarAI/sanad-agent | Skill | 47 | yesterday | MIT | A | 27 | Claude Code, Codex |
| release-changelog-harnessArenukvern/mcp_flutter | Skill | 395 | 20d ago | MIT | A | 84 | Claude Code |
| gitnexus-cliArenukvern/mcp_flutter | Skill | 395 | 20d ago | MIT | A | 59 | Claude Code |
| he-shipsgaabdu4/building-flutter-apps | Skill | 25 | 4d ago | MIT | A | 52 | Claude Code |
| code-reviewsgaabdu4/building-flutter-apps | Skill | 25 | 4d ago | MIT | A | 39 | Claude Code, Codex |
| flutterpledgeandgrow/pledge-skills | Skill | 10 | 3mo ago | MIT | A | 126 | Claude Code, Codex |
| autofix-prQwenLM/qwen-code | Skill | 29k | yesterday | Apache-2.0 | A | 141 | Claude Code |