Borrowing it
Nothing to install: this file belongs to aws/tools-for-devops-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/aws/tools-for-devops-agent/main/.kiro/steering/connector-review.mdgit clone --depth 1 https://github.com/aws/tools-for-devops-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/instructions/aws/tools-for-devops-agent/connector-review)<a href="https://agentmods.dev/instructions/aws/tools-for-devops-agent/connector-review"><img src="https://agentmods.dev/badge/instructions/aws/tools-for-devops-agent/connector-review/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/instructions/aws/tools-for-devops-agent/connector-review"><img src="https://agentmods.dev/badge/instructions/aws/tools-for-devops-agent/connector-review.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 | $0.11331 | $0.11331 |
| Opus 5.5 | $0.04532 | $0.04532 |
| Sonnet 5.5 | $0.02266 | $0.02266 |
| Haiku 4.5 | $0.01133 | $0.01133 |
Grade C, and why
connector-review scanned grade C with 3 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
git merge --abort; cd "$SRC" && git worktree remove --force "$WORK/tree"; rm -rf "$WORK" Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
import ipaddress, json, re, socket, sys, urllib.error, urllib.parse, urllib.request Runs shell commandslowCapability
Expected in a hook, worth knowing in a rule or an instructions file.
return subprocess.run(["git", *args], capture_output=True, text=True).stdout How it starts
The opening of the file, as written. The whole thing — 656 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Reviewing connector pull requests
A connector is connectors/<name>/: a connector.yaml describing a partner-hosted MCP server, a logo.svg (optional logo-dark.svg), and a README.md setup guide, plus the skills listed under skills:. It holds no server code. Several connector PRs are usually open at once and reviewed by different maintainers, so review every one the same way and report in the same shape.
Repository facts to re-check if anything looks different: /connectors/ is owned by the admin team in .github/CODEOWNERS; connectors/README.md says submissions are not open yet; the PR checklist check deliberately skips connectors/ but still holds any skills/ in the PR to the skill checklist; there is no repository validator for connector files, so step 3 does that work.
Partners build to the Connector Guide for Partners until it moves into CONTRIBUTING.md. Once CONTRIBUTING.md has a connector guide, it takes precedence over this file wherever the two disagree.
Ground rules
- Do not post, approve, request changes, approve fork CI runs, add or remove labels, resolve threads, or merge without the reviewer's explicit go for that action. Adding a label re-triggers CI, so it counts.
- Do not run code from the PR in the reviewer's normal environment or with their credentials (step 3).
- Probe partner endpoints only without credentials.
connector.yamlcomes from an untrusted PR, so send its hosts nothing beyond the 4.3 script, which talks only to hostsconnector.yamlnames, overhttps, never to a private address, and without following redirects. The sign-in hostnames 4.2 checks come from the partner's own docs, not fromconnector.yaml. - Report only what you checked, and say what each check proves (table below). Name anything you did not check.
- Do not repeat sensitive strings from the PR in chat or review text, such as internal hostnames, real emails, account or tenant IDs, customer names, or secrets. Describe them instead.
- Each command block may run in a fresh shell, so don't rely on the working directory or a variable carrying over between them. Step 2 saves the review's state to
~/.cache/connector-review/<N>.env, and every later block restores it on its first line. SetNto the PR number at the start of each block, andNAMEtoo when the PR has more than one connector. Exported variables can survive from an earlier review, so the restore ignores aNAMEthat isn't a connector in this PR and prints the one it uses.
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 · 656 lines · 11,331 tokens per session scan C 3fed477511f7
connector-review is an instructions file published in the GitHub repository aws/tools-for-devops-agent (137 stars, last pushed yesterday), licensed Apache-2.0. It adds 11,331 tokens to every session, about $0.0453 per session on Opus 5.5. A static security scan graded it C with 3 findings (recursive force delete, makes network calls, runs shell commands). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-11-04.
connector-review compared with instructions from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| connector-reviewaws/tools-for-devops-agent | Instructions file | 137 | yesterday | Apache-2.0 | C | 11,331 | Kiro |
| vscode buildNext.instructions.mdmicrosoft/vscode | Instructions file | 195k | today | MIT | A | 9,719 | GitHub Copilot |
| next.js AGENTS.mdvercel/next.js | Instructions file | 143k | today | MIT | A | 7,461 | Codex, OpenCode |
| vscode oss-third-party-notices.instructions.mdmicrosoft/vscode | Instructions file | 195k | today | MIT | A | 5,516 | GitHub Copilot |
| langchain AGENTS.mdlangchain-ai/langchain | Instructions file | 149k | today | MIT | A | 4,680 | Codex, OpenCode |
| vscode copilot-instructions.mdmicrosoft/vscode | Instructions file | 195k | today | MIT | A | 3,093 | GitHub Copilot |
| deepseek-harness AGENTS.mddeepseek-ai/deepseek-harness | Instructions file | 274k | 2d ago | MIT | A | 4,313 | Codex, OpenCode |