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 agentmods add skills/cloud-forest-analytics/evergreen/maintainnpx skills add Cloud-Forest-Analytics/evergreen --skill maintaingit clone --depth 1 https://github.com/Cloud-Forest-Analytics/evergreenWhat 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 | $0.00117 | $0.01288 |
| Opus 5 | $0.00059 | $0.00644 |
| Sonnet 5 | $0.00023 | $0.00258 |
| Haiku 4.5 | $0.00012 | $0.00129 |
Grade A, and why
maintain scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 89 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Maintain — keep the repo evergreen
GitHub is the source of truth. Every figure here is read from gh against the live board, never
from recall. When memory and GitHub disagree, GitHub wins.
One command produces the whole picture — run it first in both startup and shutdown:
"${CLAUDE_PLUGIN_ROOT}"/scripts/snapshot.sh # board + CI + DORA snapshot
"${CLAUDE_PLUGIN_ROOT}"/scripts/snapshot.sh --dora 7 # DORA over a 7-day window
Everything below interprets its output.
STARTUP (SOD) — /maintain startup
Do these in order. Do not start new work until the board is clean — finish what's in flight first.
- Ground truth.
git fetch originand confirm the primary checkout is on cleanmainsynced toorigin/main. Note themainHEAD. Don't develop on / dirty the primary branch. - CI health. Any failing or stuck runs on
main?gh run list --branch main --limit 5. A redmainis the first thing to fix. - In-flight first (the discipline that matters most).
gh pr list --state open— triage every PR: CI state, mergeable, stale. Merge anything green. Rebase anything stale. Finish near-done work (green / one fix away) before opening a single new branch. The existing backlog is the cheapest progress on the board. - Board picture. Open issues, their grouping, and the orphan-story count (open issues not linked to any active epic/milestone; target a small ceiling, e.g. <=10). Close-on-merge stragglers.
- Worktree reconcile.
"${CLAUDE_PLUGIN_ROOT}"/scripts/worktree-reconcile.sh— surface merged/closed worktrees for cleanup; zero orphans. Surface ambiguous ones, never force-remove. - Maintenance slot (anti-starvation). Confirm >=1 maintenance item is in flight. If none, start one before any feature work (see Anti-starvation below).
- Pick the focus from the roadmap's named next deliverable and start it — don't end on "standing by".
SHUTDOWN (EOD) — /maintain shutdown
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.
- yesterday First seen · 89 lines · 117 tokens per session scan A e8d7f505a268
maintain is a skill published in the GitHub repository Cloud-Forest-Analytics/evergreen (1 stars, last pushed 2mo ago), licensed MIT. It adds 117 tokens to every session and 1,288 once invoked, about $0.0006 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 2026-08-31.
Other skills, from other repositories
article-magazine
Substack / Medium 高级感长文排版, 适合公众号、博客发布.
email-marketing
产品发布邮件, 含 masthead、hero、CTA、规格表, table-fallback.
unicli-repair
Evidence-driven repair workflow for a broken Uni-CLI adapter. Trigger on a failed unicli envelope, a quarantined adapter, or an explicit adapter-repair request. Classifies non-source failures, edits only the reported adapter path, and uses the original command as a bounded oracle.
janitor-fix
Automatically fix skill problems (safe preview first). Also use with --prune to find and remove broken symlinks, empty directories, and orphaned skills. Trigger with '/janitor-fix'.
janitor-value
Show whether each skill is earning its context-window cost — combined tokens-used view sorted by waste. Use when the user asks 'are my skills worth it', 'what's my context budget', 'which skills are dead weight', or wants to audit skill value, token cost, or usage. Trigger with '/janitor-value'.
memory-cleanup
Audits all memory systems in a Perseus environment, identifies stale or superseded stores (Mem0, Mempalace, old vector DBs, garbled federation artifacts), and deletes them. Keeps Perseus Vault (the live system) as the single source of truth. Run periodically or when upgrading Perseus versions.