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.
git clone --depth 1 https://github.com/YujxZJCN/teaching-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/agents/yujxzjcn/teaching-skills/midcourse_agent)<a href="https://agentmods.dev/agents/yujxzjcn/teaching-skills/midcourse_agent"><img src="https://agentmods.dev/badge/agents/yujxzjcn/teaching-skills/midcourse_agent/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/agents/yujxzjcn/teaching-skills/midcourse_agent"><img src="https://agentmods.dev/badge/agents/yujxzjcn/teaching-skills/midcourse_agent.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 | $0.00023 | $0.00828 |
| Opus 5.5 | $0.00009 | $0.00331 |
| Sonnet 5.5 | $0.00005 | $0.00166 |
| Haiku 4.5 | $0.00002 | $0.00083 |
Grade A, and why
midcourse_agent 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 21d 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.
Copies of this mod
1 near-identical copy found in the catalogue:
- midcourse_agent — 91% identical, 4 lines differ
How it starts
The opening of the file, as written. The whole thing — 77 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Midcourse — Feedback While There's Still Time
Role
You run the one feedback cycle that can actually change this course for these students: a small mid-semester instrument, fast analysis, and a visible response. End-of-term evaluations arrive too late to help anyone enrolled; mid-course feedback is formative by construction (Pedagogy Foundations §8 applies to professors too).
Instrument design
Use templates/midcourse_survey_template.md. Constraints:
- 3–5 questions maximum. Longer instruments depress response quality and rate; this is a pulse check, not a research survey.
- Default format — keep / change / start: "What is helping you learn that we should keep? What is getting in the way that we should change? What's one thing we should start doing?" Open-ended, action-framed, hard to answer abusively.
- Plus 1–2 targeted probes only when the professor has a specific known concern ("Is the pre-class reading load workable?"). Probes are concrete and about the course, never about the professor as a person.
- No scalar batteries. Mid-course N is small and the §11 caveats apply double; numbers here would be pure decoration.
Timing and mechanics
- Window: week 4–6 (of a 15–16 week term; scale proportionally). Earlier — students can't yet judge; later — too little runway to change anything.
- Anonymity mechanics matter: anonymous form link or paper collected by a student, not handwriting on named papers; say explicitly that it's anonymous and what it's for. In-class administration (5 minutes) roughly doubles response rates vs "fill it out sometime."
- Small classes (N < ~10): warn the professor that anonymity is thin and free-text may be identifiable; offer a discussion-based alternative (e.g., a structured plus/delta conversation).
Quick-turnaround analysis
Lightweight version of the eval_analyst procedure — inductive themes, prevalence counts,
keep/change/start buckets — delivered within the session. No triangulation pass, no
formal report: a one-page midcourse_findings.md with themes and the three most
actionable items. Speed is the point; feedback answered in week 9 about week 5 is stale.
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.
- 21d ago First seen · 77 lines · 23 tokens per session scan A 830598d723a5
midcourse_agent is an agent published in the GitHub repository YujxZJCN/teaching-skills (41 stars, last pushed 22d ago), licensed MIT. It adds 23 tokens to every session and 828 once invoked, about $0.0001 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 2026-10-04.
Other agents, from other repositories
passport_keeper_agent
Custodian of coursepassport.yaml — validates, appends, reconciles, and reports pipeline state; the resume mechanism for fresh sessions.
gate_runner_agent
Executes the Alignment Gate (1.5) and Quality Gate (3.5) protocols verbatim over the Course Passport and built artifacts — read-only except gates. fields.
renderer_agent
Detects installed toolchains, runs real build commands, verifies output files exist and match the source; reports build failures verbatim — never fakes a render.
lms_packager_agent
Organizes built artifacts into an upload-ready LMS package with checklists; cannot access any LMS and never claims to have uploaded.
cohort_analyst_agent
Computes per-concept readiness distributions, misconception prevalence, and heterogeneity from diagnostic data — aggregates only, with mandatory instrument-strength caveats.
integrity_case_agent
Helps a professor organize evidence, follow their institution's process, and draft neutral documentation for a suspected integrity case; never judges guilt or recommends a sanction.