Inbox Triage logo

Inbox Triage

Organization
aeonfun
inbox-triage

Daily GitHub notification inbox triage - surfaces aging vuln PR replies, security advisories, review requests, and mentions that need action

Overview

Publisheraeonfun
Repositoryaeon
Skill nameinbox-triage
Stars
771
Forks
280
Bundled files
Instructions only
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by aeonfun on GitHub. Read the source before you install it.

Installation

Install the Inbox Triage AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/aeonfun/aeon/tree/main/skills/inbox-triage
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/aeonfun/aeon.git /tmp/aeon
mkdir -p .claude/skills
cp -r /tmp/aeon/skills/inbox-triage .claude/skills/inbox-triage
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Inbox Triage in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Inbox Triage on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Inbox Triage is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Today is ${today}. Read memory/MEMORY.md before starting.

Why this skill exists

Hand-tracked follow-ups live in MEMORY.md. vuln-tracker tracks the operator's vuln PRs by scanning branch names and ages the memory/pending-disclosures/ queue. Neither reads from the actual GitHub notification inbox. When a maintainer replies to a vuln PR — or a security advisory opens on a watched repo — it sits unread until someone manually checks GitHub. This skill reads the inbox and routes what needs action.

vuln-tracker covers PR and disclosure lifecycle by branch and draft state. This skill covers the notification layer — inbound responses, review requests, security alerts, mentions.

Steps

1. Fetch GitHub notifications

Run:

bash
gh api /notifications --paginate 2>&1

Parse the JSON array. If the command errors or returns an empty array [], log INBOX_TRIAGE_SKIP: no notifications and stop.

Limit to the first 100 notifications if --paginate returns more (GitHub caps at 50 per page; two pages is enough).

For each notification record:

  • id
  • reason — why you're being notified (mention, review_requested, author, state_change, security_alert, assign, etc.)
  • subject.title
  • subject.type — PullRequest, Issue, Release, etc.
  • subject.url — API URL for the subject
  • repository.full_name
  • updated_at — ISO timestamp

2. Filter

Keep notifications where unread: true AND updated_at is within the last 14 days. Discard older or read ones.

If zero remain after filtering: log INBOX_TRIAGE_SKIP: no actionable notifications within 14 days and stop.

3. Categorize

Assign each notification to exactly one category (first match wins):

CategoryMatch criteria
SECURITYreason == "security_alert" OR title contains any of: vulnerability, vuln, CVE, advisory, security
VULN_REPLYsubject.type == "PullRequest" AND reason is one of: author, state_change, comment AND repository.full_name is NOT under the operator's own account/org (derive the operator's GitHub handle from soul/SOUL.md or the workflow's GITHUB_ACTOR — these are PRs filed on third-party repos by the vuln-scanner)
REVIEW_NEEDEDreason == "review_requested"
MENTIONreason == "mention" OR reason == "team_mention"
GENERALeverything else

4. Age vuln PR replies

For each VULN_REPLY notification, compute age_days = today minus updated_at date (integer days).

Flag urgency:

  • CRITICAL — age_days > 7 (maintainer likely hasn't responded)
  • AGING — age_days 3–7
  • FRESH — age_days < 3

Cross-reference with memory/topics/vuln-followup.md if it exists: look for the PR title in that file and pull any tracked notes (e.g. "approved", "NEEDS-ANSWER", merge status).

5. Resolve HTML URLs for action items

For each notification in SECURITY, VULN_REPLY (CRITICAL or AGING), REVIEW_NEEDED, and MENTION categories:

Try to get the HTML URL via:

bash
gh api {subject.url} --jq '.html_url' 2>/dev/null

If that fails, construct the URL manually: https://github.com/{repository.full_name}/pulls/{number} for PRs https://github.com/{repository.full_name}/issues/{number} for issues

(Extract the number from the tail of subject.url.)

6. Write triage summary

Overwrite memory/topics/inbox-triage.md:

markdown
# GitHub Inbox Triage

Last run: {today}
Scanned: {N} unread notifications ({N} within 14 days)

## Action Required

### Security ({count})
{for each SECURITY item, sorted by age:}
- **{repo}**: {title} ({age_days}d) — {html_url}

{if none:}
None.

### Vuln PR Replies ({count_critical} critical, {count_aging} aging)
{for each VULN_REPLY sorted by age desc:}
- **[{CRITICAL|AGING|FRESH}]** `{repo}` ({age_days}d): {title} — {html_url}
  {if vuln-followup note found:} _{tracked note}_

{if none:}
None.

### Review Requested ({count})
{for each REVIEW_NEEDED item:}
- **{repo}**: {title} — {html_url}

{if none:}
None.

### Mentions ({count})
{for each MENTION item:}
- **{repo}**: {title} — {html_url}

{if none:}
None.

## No Action Needed
{count_general} general notifications (subscriptions, automated state changes).

7. Update MEMORY.md known follow-ups

Read memory/MEMORY.md. Find the ## Known Follow-ups section.

Add any VULN_REPLY CRITICAL item not already tracked there — append:

- **{repo} #{number} NEEDS-ANSWER** — {age_days}d since maintainer activity ({url})

Update any existing NEEDS-ANSWER item for a PR that now appears as FRESH in VULN_REPLY (maintainer responded recently) — change its note to RESPONDED — verify resolution.

Do NOT add GENERAL, REVIEW_NEEDED, MENTION, or SECURITY items to MEMORY.md Known Follow-ups (too noisy; security items warrant a separate issue if severe).

8. Send notification

Only send if at least one of:

  • Any SECURITY item
  • Any VULN_REPLY where urgency == CRITICAL
  • Any REVIEW_NEEDED item
  • Three or more MENTION items

Write to .pending-notify-temp/inbox-triage-${today}.md:

inbox — {today}

{if SECURITY:}
security alert: {repo} — {title}

{if VULN_REPLY CRITICAL:}
vuln PRs aging: {comma-separated list of "repo (Nd)"}

{if REVIEW_NEEDED:}
review needed: {comma-separated repo list}

{if 3+ MENTION:}
{N} mentions

read it: memory/topics/inbox-triage.md

Then:

bash
./notify -f .pending-notify-temp/inbox-triage-${today}.md

If nothing meets the threshold: skip notification. Log that no notification was sent.

9. Log

Append to memory/logs/${today}.md:

markdown
### inbox-triage
- **Scanned:** {N} notifications
- **Security:** {N}
- **Vuln replies:** {N total} ({N_critical} critical, {N_aging} aging, {N_fresh} fresh)
- **Review needed:** {N}
- **Mentions:** {N}
- **MEMORY.md follow-ups updated:** {yes/no — what changed}
- **Notification sent:** {yes/no}
- INBOX_TRIAGE_OK

If skipped:

markdown
### inbox-triage
- INBOX_TRIAGE_SKIP: {reason}

Required Env Vars

None beyond GITHUB_TOKEN, which GitHub Actions sets automatically and gh uses internally.

Network Note

Uses gh api for all GitHub calls — it handles auth internally, so no $SECRET ever appears on the command line for the Bash permission layer to refuse. gh api works in a GitHub Actions run. If gh api /notifications fails (rate limit, auth error), log the error and exit with INBOX_TRIAGE_SKIP: api error. Use WebFetch as a fallback only if gh is unavailable — the endpoint is https://api.github.com/notifications with Authorization: Bearer $GITHUB_TOKEN, but WebFetch can't carry that auth header; prefer gh.

What this is NOT

  • Not a replacement for hand-tracked follow-ups in MEMORY.md. This reads the raw GitHub inbox.
  • Not a duplicate of vuln-tracker — vuln-tracker tracks lifecycle by branch name. This catches inbound maintainer replies via notifications.
  • Not a duplicate of vuln-tracker's disclosure-queue aging - that manages memory/pending-disclosures/ advisory drafts. This reads GitHub security alerts and PR responses.

Frequently asked questions

What does the Inbox Triage AI skill do?

Daily GitHub notification inbox triage - surfaces aging vuln PR replies, security advisories, review requests, and mentions that need action

Why use Inbox Triage on TypingMind?

Because you install it once and use it with any model. Inbox Triage is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Inbox Triage in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/aeonfun/aeon/tree/main/skills/inbox-triage. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Inbox Triage?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Inbox Triage?

As many as you like. As long as a model supports skills, you can use Inbox Triage with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Inbox Triage AI skill free?

Yes. It is published on GitHub by aeonfun under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇