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 wiber/thetacog-mcp --skill allocategit clone --depth 1 https://github.com/wiber/thetacog-mcpWrote 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/wiber/thetacog-mcp/allocate)<a href="https://agentmods.dev/skills/wiber/thetacog-mcp/allocate"><img src="https://agentmods.dev/badge/skills/wiber/thetacog-mcp/allocate/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/wiber/thetacog-mcp/allocate"><img src="https://agentmods.dev/badge/skills/wiber/thetacog-mcp/allocate.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 | $0.00092 | $0.01098 |
| Opus 5.5 | $0.00037 | $0.00439 |
| Sonnet 5.5 | $0.00018 | $0.00220 |
| Haiku 4.5 | $0.00009 | $0.00110 |
Grade A, and why
allocate 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 — 67 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/allocate — one session, one spec, is the default
CC apps usually run one session at a time. The default path here matches that: one
Claude Code session, one spec, every job in the ask becomes a row on that spec, and
/steer works the rows one at a time through the installed doors (npx thetacog-mcp <door>), never a thetadrivencoach-only command. The chair reasons; the dispatched
claude -p workers build. This keeps the one session's own context clean — it holds
the map and the next row, not every row's history at once.
The default: one session, one spec, rows in order
- Split the ask into rows on ONE spec (
npx thetacog-mcp steer-compile "<the ask>"does this: one row per job, a guard proposed for each). - Give each row one area — a row that touches two unrelated areas gets split again.
- Run the rows through
/steerone at a time. Check the goal before starting the next row (npx thetacog-mcp goal): a row still open stays open until its own guard is green, never folded into the row after it. - Check who is working and end anything that should stop before starting more
(
npx thetacog-mcp workers list).
The doors
| what it does | door |
|---|---|
| split an ask into rows on one spec, one guard proposed per row | npx thetacog-mcp steer-compile "<ask>" --check (status only; drop --check to run) |
| the goal line — which row is open, which is done | npx thetacog-mcp goal |
| who is working right now — list, or end one | npx thetacog-mcp workers list |
| keep the one-row-at-a-time loop running — on, off, status | npx thetacog-mcp keep-running |
Every one of these is a door the extension itself dispatches through (npx thetacog-mcp --help prints the live list). A job that needs a door not on this list
is not runnable here yet — say so plainly rather than inventing a path.
Appendix: several sessions at once (optional, never the default)
Running more than one Claude Code session on the same working tree is possible. Each lesson below cost something the night this skill was written; each names the failure, then the rule.
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 · 67 lines · 92 tokens per session scan A 5c65303ea509
allocate is a skill published in the GitHub repository wiber/thetacog-mcp (1 stars, last pushed yesterday), licensed MIT. It adds 92 tokens to every session and 1,098 once invoked, about $0.0004 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 2026-11-06.
allocate compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| allocatewiber/thetacog-mcp | Skill | 1 | yesterday | MIT | A | 92 | Claude Code, Codex |
| git-checkpointagentscope-ai/agentscope-java | Skill | 6.3k | today | Apache-2.0 | A | 39 | Claude Code, Codex |
| synthesize-reportagentscope-ai/agentscope-java | Skill | 6.3k | today | Apache-2.0 | A | 29 | Claude Code, Codex |
| data-analysisagentscope-ai/agentscope-java | Skill | 6.3k | today | Apache-2.0 | A | 25 | Claude Code, Codex |
| summarizeragentscope-ai/agentscope-java | Skill | 6.3k | today | Apache-2.0 | A | 24 | Claude Code, Codex |
| agent-browserTencent/WeKnora | Skill | 37k | 4d ago | none | A | 83 | Claude Code, Codex |
| skill-creator-primersammcj/agentic-coding | Skill | 169 | 3d ago | Apache-2.0 | A | 74 | Claude Code |