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 skills add HubSpot/agent-cli-skills --skill conversation-inbox-triagegit clone --depth 1 https://github.com/HubSpot/agent-cli-skillsWrote 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/hubspot/agent-cli-skills/conversation-inbox-triage)<a href="https://agentmods.dev/skills/hubspot/agent-cli-skills/conversation-inbox-triage"><img src="https://agentmods.dev/badge/skills/hubspot/agent-cli-skills/conversation-inbox-triage/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/hubspot/agent-cli-skills/conversation-inbox-triage"><img src="https://agentmods.dev/badge/skills/hubspot/agent-cli-skills/conversation-inbox-triage.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 | $0.00048 | $0.03220 |
| Opus 5.5 | $0.00019 | $0.01288 |
| Sonnet 5.5 | $0.00010 | $0.00644 |
| Haiku 4.5 | $0.00005 | $0.00322 |
Grade A, and why
conversation-inbox-triage 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 1mo 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 — 275 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Requires the conversations command group from hub-cli PR #251. hubspot conversations --help is authoritative. Archive uses the dry-run/digest/confirm guard pattern — follow it exactly.
Paging: never triage from a --format jsonl redirect
Every conversations list command defaults to --limit 100 and returns a cursor for the next page. --format jsonl prints only the data rows — the cursor is surfaced only by --format json, at .meta.next. So threads list --format jsonl > file silently stops at 100 threads: on a busy inbox the operator sees a clean run while every thread past the first page is never assigned, never archived, and missing from the counts. Raising --limit does not fix it — the endpoint caps the page size.
Use this helper for every bulk or counting operation:
# conv_page <output-file> <conversations list args...>
# Pages a conversations list endpoint to exhaustion, writing one JSON object
# per line to <output-file>.
conv_page() {
out=$1; shift
after=""
: > "$out"
while :; do
if [ -n "$after" ]; then
page=$(hubspot "$@" --format json --after "$after") || return 1
else
page=$(hubspot "$@" --format json) || return 1
fi
printf '%s' "$page" | jq -c '.data[]' >> "$out"
after=$(printf '%s' "$page" | jq -r '.meta.next // empty')
[ -n "$after" ] || break
done
}
The bare --format jsonl calls in steps 2–4 are fine for eyeballing the first page interactively. Steps 6–8 mutate or count, so they page.
1. Discover inboxes
Inbox IDs are portal-specific. Always enumerate at the start of a session — never hardcode.
conv_page /tmp/inboxes.jsonl conversations inboxes list
# {"id":"123","name":"Support Inbox","createdAt":"...","updatedAt":"..."}
# {"id":"124","name":"Sales Chat","createdAt":"...","updatedAt":"..."}
To see channels inside an inbox (email, live chat, form, etc.):
conv_page /tmp/channels.jsonl conversations channels list --inbox-id 123
# {"id":"456","name":"Support Email","type":"EMAIL","inboxId":"123"}
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.
- 1mo ago First seen · 275 lines · 48 tokens per session scan A b7fc35d343c3
conversation-inbox-triage is a skill published in the GitHub repository HubSpot/agent-cli-skills (28 stars, last pushed 9d ago), licensed Apache-2.0. It adds 48 tokens to every session and 3,220 once invoked, about $0.0002 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 2026-09-26.
conversation-inbox-triage compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| conversation-inbox-triageHubSpot/agent-cli-skills | Skill | 28 | 9d ago | Apache-2.0 | A | 48 | Claude Code, Codex |
| imagegenopenai/codex | Skill | 133k | yesterday | Apache-2.0 | A | 113 | Codex |
| playwright-climicrosoft/playwright | Skill | 99k | yesterday | Apache-2.0 | A | 19 | Claude Code |
| local-ai-agentsmicrosoft/ai-agents-for-beginners | Skill | 79k | 4d ago | MIT | A | 200 | Claude Code, Codex |
| systematic-debuggingobra/superpowers | Skill | 313k | 3d ago | MIT | A | 21 | Claude Code |
| next-cache-components-adoptionvercel/next.js | Skill | 143k | yesterday | MIT | A | 95 | Claude Code |
| next-partial-prefetching-adoptionvercel/next.js | Skill | 143k | yesterday | MIT | A | 103 | Claude Code |