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/bradfeld/whats-new-pluginWrote 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/commands/bradfeld/whats-new-plugin/whats-new)<a href="https://agentmods.dev/commands/bradfeld/whats-new-plugin/whats-new"><img src="https://agentmods.dev/badge/commands/bradfeld/whats-new-plugin/whats-new/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/commands/bradfeld/whats-new-plugin/whats-new"><img src="https://agentmods.dev/badge/commands/bradfeld/whats-new-plugin/whats-new.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 | $0.00009 | $0.04411 |
| Opus 5 | $0.00005 | $0.02205 |
| Sonnet 5 | $0.00002 | $0.00882 |
| Haiku 4.5 | $0.00001 | $0.00441 |
Grade B, and why
whats-new scanned grade B with 2 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 10d 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.
Reads agent configuration directoriesmediumAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
| "sandbox" + "worktree" / "main repo" / "shared `.git`" | Run `git worktree list` to enumerate worktrees. `grep -rln "<main-repo-name>" ~/.claude/hooks/ ~/.claude/scripts/` and inspect for writes targeting the main repo Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
curl -s "https://api.github.com/repos/anthropics/claude-code/releases?per_page=100" How it starts
The opening of the file, as written. The whole thing — 316 lines — stays where its author put it; the contents beside it link to each section on GitHub.
What's New
Help Mode
If the argument is ? or --help, print this block and STOP — do not execute any steps:
Cross-reference Claude Code release notes against your personal config,
then verify each impact item with concrete tool calls before reporting.
Surfaces what changed that actually requires action, recommends new
features worth adopting, and summarizes the rest.
Usage: /whats-new [version]
Arguments:
(none) Analyze all releases since last reviewed
<version> Analyze a specific version only (e.g., 2.1.83)
Does not update the last-reviewed tracking.
Examples:
/whats-new Full analysis since last review
/whats-new 2.1.83 Analyze a specific version
/whats-new ? Show this help
Each IMPACT item is verified with grep / read / file-check before being
reported as actionable. Items that the verification proves don't affect
your config are demoted to GENERAL — no false alarms.
In since-last-review mode, after the report it independently reviews each
derived config edit, then proposes the vetted changes as a numbered list
(each with a plain-language explanation and the review's verdict); reply
with the numbers to apply, "all", or "none". Applied changes are logged to
~/.claude/whats-new-applied.md.
Step 1 — Parse Arguments
Read $ARGUMENTS:
- If empty: set
mode = since-last-review - If the value matches a semver pattern like
2.1.83(digits, dot, digits, dot, digits): setmode = specific-version, settarget_version= that value - If
?or--help: already handled by Help Mode above — stop here
Step 2 — Get Current Version
Use Bash tool to run:
claude --version | grep -oE '[0-9]+\.[0-9]+\.[0-9]+'
Store the result as current_version.
Step 3 — Determine Version Range
If mode = since-last-review:
- Use the Read tool to read
~/.claude/whats-new-last-version.txt - If the file exists and has content: set
from_version= file contents (trimmed of whitespace) - If the file is missing or empty: skip to Step 4 to fetch the releases list first, then:
- Find
current_versionin the newest-first release list - Set
from_version= the version 5 positions further into the list (i.e.,releases[current_idx + 5]) - If
current_versionis not in the list, use the most recent published release asfrom_version
- Find
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.
- 10d ago First seen · 316 lines · 9 tokens per session scan B 7f2e473bcb26
whats-new is a command published in the GitHub repository bradfeld/whats-new-plugin (10 stars, last pushed 1mo ago), licensed MIT. It adds 9 tokens to every session and 4,411 once invoked, about $0.0000 per session on Opus 5. A static security scan graded it B with 2 findings (reads agent configuration directories, makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other commands, from other repositories
checklist
Generate a custom checklist for the current feature based on user requirements.
clarify
Identify underspecified areas in the current feature spec by asking up to 5 highly targeted clarification questions and encoding answers back into the spec.
specify
Create or update the feature specification from a natural language feature description.
analyze
Perform a non-destructive cross-artifact consistency and quality analysis across spec.md, plan.md, and tasks.md after task generation.
converge
Assess the current codebase against the feature's spec, plan, and tasks, then append any remaining unbuilt work as new tasks to tasks.md so implement can complete it.
implement
Execute the implementation plan by processing and executing all tasks defined in tasks.md.