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 huangpufan/SSOT-SKILL --skill ssot-closeoutgit clone --depth 1 https://github.com/huangpufan/SSOT-SKILLWrote 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/huangpufan/ssot-skill/ssot-closeout)<a href="https://agentmods.dev/skills/huangpufan/ssot-skill/ssot-closeout"><img src="https://agentmods.dev/badge/skills/huangpufan/ssot-skill/ssot-closeout/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/huangpufan/ssot-skill/ssot-closeout"><img src="https://agentmods.dev/badge/skills/huangpufan/ssot-skill/ssot-closeout.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 | $0.00087 | $0.01338 |
| Opus 5.5 | $0.00035 | $0.00535 |
| Sonnet 5.5 | $0.00017 | $0.00268 |
| Haiku 4.5 | $0.00009 | $0.00134 |
Grade A, and why
ssot-closeout scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured today.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 41 lines — stays where its author put it; the contents beside it link to each section on GitHub.
SSOT Closeout
Bring SSOT/ in line with what this batch changed: the diff, the decisions made in the conversation, and what you observed while working. What to store, where it lives, and how to write it are defined in memory-model.md.
Before you start
- No
SSOT/directory. Nothing to do, unless the user asked to create one (that is$ssot-bootstrap). - Not v3 (
SSOT/README.mdhas nossot_format: 3, usually a v2 layout withSTATUS.md). Do not write. List the durable facts from this batch for the user, explain that SSOT stays read-only until it is migrated, and offer the migration with$ssot-bootstrap. - Setup unfinished (
SSOT/BOOTSTRAP.mdexists). Update documents that already exist. Note anything else inBOOTSTRAP.mdso setup picks it up.
The batch is the work done in this task. If other people's changes are in the tree, or HEAD moved underneath you, do not record facts from them that you have not verified. Tell the user, and leave them to a $ssot-doctor catch-up.
Steps
-
List the candidates. Go through the batch and look for:
- decisions made, and alternatives the user rejected;
- contracts, APIs, data formats, or behavior that users rely on and that changed;
- traps you hit, with their cause;
- build, test, release, or deploy commands that changed;
- temporary workarounds added, with their removal condition;
- lasting directives from the user;
- findings that will keep guiding work: a failed approach, an unexplained observation, a research result;
- SSOT statements this work proved wrong;
- protected rules found during the work that SSOT lacks (preflight section 3), as candidates.
Before you decide there is nothing, open the owners that describe the parts this batch touched, plus any related problems preflight surfaced, and compare them with what is now true. A statement that is now wrong must be fixed or removed, even if the fact behind it could be recovered from the code. "Could be recovered from the code" only decides what to add, never what to leave wrong. If the owners still hold, write nothing: no logs, no timestamps, no "checked" markers. Say so in your report. If you already updated the owners during the work, that counts.
-
Write each one in its single owner. Use the README's routes and the model's layout. Update in place, in the present tense. Add a decision record for a consequential choice. Record a lasting user directive, or a departure from a protected rule that the user accepted, as its own confirmed row: its Source quotes the user's words and the date, and the row states only what those words decided. Clauses of an earlier candidate row that the user did not discuss stay in a candidate row; never move a clause of a confirmed row to candidate. Put the explanation in the owner the row links to. A one-time choice the user made goes to a decision record instead. Label anything decided but not built as planned.
-
Remove what it supersedes. Search
SSOT/and the agent instruction files (AGENTS.md,CLAUDE.md) for older statements of the same fact (names, numbers, the old behavior) that still act as current guidance, and fix or delete them. A rule kept short in an instruction file and explained inSSOT/changes in both places. If you renamed or split a file or heading, fix the links to it across the repository. History keeps its meaning: in a past decision or research note, leave the old behavior as written, set a replaced decision's status tosuperseded by NNNN, and link the successor. -
Protect people's rules. Change a confirmed invariant, or an accepted decision a person made, only on an explicit user decision. A decision in this conversation counts, and so does an earlier one that is already recorded and applies. Record it with its date. Without one, pause only the work that depends on the change, and ask. Record a departure from a protected rule as decided only when the approval rule in memory-model §5 is met, as step 2 says; otherwise report the conflict to the user, and do not record the new behavior as the truth.
-
Check. Unless this batch only edited the contents of files outside
SSOT/(no file or heading added, moved, renamed, or deleted), run the checker that ships withssot-doctor, passing the commit the batch started from (HEADif nothing is committed yet):python3 <ssot-doctor skill directory>/assets/scripts/ssot-check.py SSOT --since <commit>. Fix the errors this batch introduced, plus any that affect what you wrote. A warning that a file this batch removed or moved is still mentioned points at a sentence that may now describe code that is gone: correct it if it does. Report other errors without fixing them, and do not present a scoped check as a clean bill for the whole SSOT. Then reread each changed passage as a newcomer would. -
Report. In your final response, name the SSOT files you changed and why. If you changed nothing, name the owners you compared and say why nothing in them became wrong or missing.
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- today Changed 06688f6875a1
- 7d ago Changed · +1 lines 45bb99c567db
- 9d ago Changed · -54 lines 3e1afa7577a0
- 10d ago Changed · +78 lines · +7 tokens per session aa0cc4128ba2
- 1mo ago Changed · +2 lines 23c87eddc83d
- 2mo ago First seen · 14 lines · 80 tokens per session scan A 9a42f73320f7
ssot-closeout is a skill published in the GitHub repository huangpufan/SSOT-SKILL (11 stars, last pushed today), licensed MIT. It adds 87 tokens to every session and 1,338 once invoked, about $0.0003 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 2026-08-30.
ssot-closeout compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| ssot-closeouthuangpufan/SSOT-SKILL | Skill | 11 | today | MIT | A | 87 | 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 | today | MIT | A | 95 | Claude Code |
| next-partial-prefetching-adoptionvercel/next.js | Skill | 143k | today | MIT | A | 103 | Claude Code |