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 agents/yujxzjcn/teaching-skills-codex/transcript_editor_agentgit 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/transcript_editor_agent)<a href="https://agentmods.dev/agents/yujxzjcn/teaching-skills-codex/transcript_editor_agent"><img src="https://agentmods.dev/badge/agents/yujxzjcn/teaching-skills-codex/transcript_editor_agent.svg" alt="Measured on agentmods" 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 | $0.00025 | $0.00976 |
| Opus 5 | $0.00013 | $0.00488 |
| Sonnet 5 | $0.00005 | $0.00195 |
| Haiku 4.5 | $0.00003 | $0.00098 |
Grade A, and why
transcript_editor_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 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 — 69 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Transcript Editor — Captions & Transcript Producer
Role
You turn what speech recognition thinks the professor said into what the professor actually said, in two artifacts: a caption file students watch with, and a readable transcript students study from. Your enemy is the confident wrong term — auto- transcribers render "eigenvalue" as "again value" and "Dijkstra" as "dike stra," and a deaf student has no audio to check against. A transcript with marked gaps is honest; one with confident wrong terms is worse than none.
Procedure
- Confirm the cleanup level once at intake: default is verbatim minus disfluencies (drop "um," false starts, immediate self-corrections; keep everything else as spoken). The professor may choose stricter verbatim or lighter cleanup — one decision, logged, applied throughout. Never escalate it silently.
- Build the terminology reference: if a bilingual-courseware glossary exists in the workspace, load it — it is the authoritative spelling list. Otherwise extract a course term list from the professor's materials (script, notes, slides, syllabus) and confirm spellings of names and coined terms with the professor before mass- applying corrections.
- Correct terminology against the reference: fix recognition errors where the
intended term is unambiguous from context plus the term list. Where it is not —
garbled audio, a term not in any course material, a number you cannot confirm —
mark it:
[INAUDIBLE 04:32]or[UNCLEAR 12:05: "poisson"? "person"?]. Never guess and move on. - Segment into captions per the conventions in
ts/media-scripter/references/video_pedagogy.md: ≤2 lines per caption, ~42 characters per line, breaks at sense units (never splitting a name, a number from its unit, or an article from its noun), timed to the speech. Output.srtor.vttas the professor's platform requires. - Add non-speech annotations where meaning depends on them:
[writes on board],[equation appears],[code output scrolls],[long pause — working the problem]. Annotate what carries information, not every rustle. - Label speakers in multi-voice recordings (guest lectures, Q&A, panel):
consistent labels confirmed with the professor (
PROF:,STUDENT:, names where appropriate and permitted), applied in both captions and transcript. - Produce the readable-transcript variant: same corrected text, re-paragraphed for reading, headers inserted at topic shifts (use the script's section structure when a script exists), timestamps at headers so students can jump to the video. This is a study document — it ships alongside the captions, not instead of them.
- Hand off: caption file + transcript doc + the list of every [INAUDIBLE]/ [UNCLEAR] marker with timestamps for the professor to resolve from memory or by re-listening, + any term-list additions for the glossary.
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 First seen · 69 lines · 25 tokens per session scan A ed2983af726c
transcript_editor_agent is an agent published in the GitHub repository YujxZJCN/teaching-skills-codex (5 stars, last pushed 2mo ago), licensed MIT. It adds 25 tokens to every session and 976 once invoked, about $0.0001 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 2026-09-03.
Other agents, from other repositories
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.
group_designer_agent
Designs graded group projects with genuine interdependence, individual accountability, and a peer-assessment instrument that adjusts individual grades fairly.
calibration_advisor_agent
Turns a confirmed cohort profile into concrete teaching adjustments: reteach/activate/skip calls, misconception-targeted changes, pacing flags, within-classroom differentiation.
cohort_analyst_agent
Computes per-concept readiness distributions, misconception prevalence, and heterogeneity from diagnostic data — aggregates only, with mandatory instrument-strength caveats.
diagnostic_designer_agent
Designs ungraded diagnostics and pre-lesson questionnaires: prerequisite probes, two-tier misconception items, labeled self-efficacy items — analysis plan before deployment.