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 aozyildirim/Agena --skill stale-ticket-triagegit clone --depth 1 https://github.com/aozyildirim/AgenaWrote 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/aozyildirim/agena/stale-ticket-triage)<a href="https://agentmods.dev/skills/aozyildirim/agena/stale-ticket-triage"><img src="https://agentmods.dev/badge/skills/aozyildirim/agena/stale-ticket-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/aozyildirim/agena/stale-ticket-triage"><img src="https://agentmods.dev/badge/skills/aozyildirim/agena/stale-ticket-triage.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00057 | $0.00710 |
| Opus 5 | $0.00028 | $0.00355 |
| Sonnet 5 | $0.00011 | $0.00142 |
| Haiku 4.5 | $0.00006 | $0.00071 |
Grade A, and why
stale-ticket-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 12d 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 — 68 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Stale Ticket Triage
Backlogs accumulate dead rows. Most teams have a recurring Friday meeting where someone walks every ticket older than X days and decides whether to close, snooze, or keep it. That hour is the perfect AI workload — the decision is shallow but the volume is high.
How to apply this pattern
-
Define stale: a ticket is stale when its
updated_atis older than the org's threshold (default 30 days) AND its status is still in the active set (not Closed / Done / Cancelled). -
Schedule the scan (chip-pick: every 6h / 12h / daily 9am / weekly Sundays / monthly). Each run hits the source-platform API for every project the org cares about, lists stale issues, and sends each through a short LLM call.
-
System prompt the LLM with three explicit verdicts —
close/snooze/keep— plus a one-sentence reason. Be conservative onclose: only pick it when the ticket itself signals resolution (links a merged PR, mentions a follow-up). Default tosnoozewhen unsure; neverclosefrom silence alone. -
Persist each verdict as a triage decision row keyed by
(org, source, external_id)so re-runs are idempotent. Status transitions:pending→applied/skipped/overridden. -
One-click bulk approve in the UI: the human reviews the AI verdicts in a list, hits "Apply all AI suggestions", and the system writes back to Jira / Azure DevOps in a single batch.
Example LLM reply format
VERDICT: close
REASON: Closed by PR #4221 (merged 38 days ago); customer hasn't
responded since the fix shipped.
VERDICT: snooze
REASON: Still relevant but no recent customer activity; revisit when
the related epic resumes.
Notes
- Audit trail is non-negotiable. Every AI verdict and every human override needs a row with timestamp + user id. Bulk-approve must not destroy history; it appends.
- Source-side scan (read tickets directly from Jira / Azure DevOps) scales better than relying on imported task records, because most stale tickets were never imported into the AI tool.
- Threshold is per-workspace. Solo product teams might want 14 days; enterprise backlogs need 60-90.
- The verdict prompt is short on purpose — long prompts encourage the model to add commentary and skip the structured reply.
- Pair with reporter-routing rules so security tickets bypass the triage flow entirely — those should never be auto-closed.
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.
- 12d ago First seen · 68 lines · 57 tokens per session scan A 101b8f41e3fb
stale-ticket-triage is a skill published in the GitHub repository aozyildirim/Agena (99 stars, last pushed 1mo ago), licensed MIT. It adds 57 tokens to every session and 710 once invoked, about $0.0003 per session on Opus 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.
Other skills, from other repositories
gsd-tools
Central utility skill for GSD operations. Provides config parsing, slug generation, timestamps, path operations, and orchestrates calls to other specialized skills. Acts as the unified entry point that the original gsd-tools.cjs provided via its lib/ modules (commands, config, core, init).
babysit-babysitter-issues
This skill should be used when the user asks to "babysit issues", "work on assigned issues", "check a5c-agent issues", "process babysitter issues", or wants to find and work on open GitHub issues assigned to a5c-agent in the babysitter repo.
cog-meeting-processing
Process meeting recordings and transcripts into decisions, action items, and team dynamics.
cog-team-intelligence
Cross-reference GitHub, Linear, Slack, and PostHog with bidirectional sync for team briefs.
spec-driven-development
Specification creation and management for the Pilot Shell methodology. Covers semantic search, clarifying questions, structured spec generation, and iterative refinement.
cog-onboarding
Personalize COG Second Brain workflow through role pack selection and vault initialization.