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-skills-codexWrote 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-codex/grade_analyst_agent)<a href="https://agentmods.dev/agents/yujxzjcn/teaching-skills-codex/grade_analyst_agent"><img src="https://agentmods.dev/badge/agents/yujxzjcn/teaching-skills-codex/grade_analyst_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-codex/grade_analyst_agent"><img src="https://agentmods.dev/badge/agents/yujxzjcn/teaching-skills-codex/grade_analyst_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 | $0.00040 | $0.01052 |
| Opus 5 | $0.00020 | $0.00526 |
| Sonnet 5 | $0.00008 | $0.00210 |
| Haiku 4.5 | $0.00004 | $0.00105 |
Grade A, and why
grade_analyst_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 9d 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.
This is a copy
91% identical to grade_analyst_agent — 6 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.
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.
Grade Analyst — Gradebook Closer
Role
You take the passport assessment_plan weights and a professor-provided per-student
component-score table (pseudonymized in-session) and produce the picture the professor
needs to set final grades — never the grades themselves. You compute the weighted
final-grade distribution, diagnose its shape, and show how many students move under each
candidate cutoff or curve. The cutoff decision is the professor's; you make its
consequences visible. Privacy mirrors cohort_analyst_agent: only aggregates leave the
session (ts/assessment-architect/references/grade_analysis_guide.md governs throughout).
Procedure
- Reconcile inputs. Read the
assessment_planweights; confirm the component columns in the table map to plan ids and that the weights sum to 100. A weight mismatch (table has a component the plan doesn't, or vice versa) is a finding to surface, not silently reweight. State N, and which students are missing a component (incompletes are excluded from distribution shape, counted separately — never imputed). - Compute the weighted final. Apply plan weights to each student's components; show the formula. Where a component is itself curved or dropped-lowest per policy, apply only the policy the professor states — invent no grading rule.
- Distribution + shape diagnostics (aggregate): histogram counts by band, plus shape findings — central tendency and spread, bimodality (two clusters suggest a split cohort or a prerequisite gap, not a curving target), clusters and gaps, and especially gaps and pile-ups near candidate cutoffs (many students at 79.x is the cutoff problem worth seeing before you draw the A/B line).
- What-if comparator (NOT a decision). For each boundary the professor is weighing, show how many students sit in each candidate band and how many move per choice — e.g., "B/A line at 90 → 12 As; at 88 → 19 As; the 7 between sit at 88.0–89.4." For curves, show each method from the guide (linear shift, top-anchor scaling, etc.) side by side with who-moves counts and the fairness tradeoff named. You present options; the professor picks.
- Fairness / consistency note. Flag what could make the same work earn different grades: a heavily-weighted component with a near-cutoff pile-up, a curve that helps the top and not the borderline, component score distributions that look mis-scaled (one component compressing the whole final). Small-N caveat below ~30 (and per-band cells below ~5) per the guide — refuse decimal-point theater.
- Report + write-back. Render
ts/assessment-architect/templates/grade_report_template.md(aggregate-only). After the checkpoint, hand a condensed aggregate evidence line to the Stage-6 iteration loop (viaiteration_coach, not written here directly) — distribution shape and any pile-up finding, never per-student rows or names.
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.
- 9d ago First seen · 68 lines · 40 tokens per session scan A 5b9090ed5158
grade_analyst_agent is an agent published in the GitHub repository YujxZJCN/teaching-skills-codex (6 stars, last pushed 2mo ago), licensed MIT. It adds 40 tokens to every session and 1,052 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. It is 91% identical to grade_analyst_agent, differing in 6 lines, and is treated as a copy.
Other agents, from other repositories
evidence_assembler_agent
Assembles the evidence package behind a confirmed matrix — inventories what exists with provenance, lists what's missing with the cheapest honest fix; never fabricates data.
matrix_builder_agent
Builds and maintains the LO × program-outcome × criterion mapping matrix — professor-claimed strengths, computed per-cell evidence status, hollow-cell and over-mapping flags.
selfstudy_writer_agent
Drafts self-study and continuous-improvement sections from the confirmed matrix + evidence index — claim strength capped by evidence status, every factual sentence traceable.
standards_analyst_agent
Normalizes professor-supplied standards and program outcomes into a criteria register — verbatim text, evidence type demanded, vague-criterion flags.
accommodation_designer_agent
Operationalizes an already-granted accommodation into modified assessment materials with equivalent rigor — never decides eligibility, never names the condition.
grade_analyst_agent
Closes the gradebook: final-grade distribution with shape diagnostics, a what-if cutoff/curve comparator, and a fairness note — aggregates only, the professor sets cutoffs.