Skills

Appearance

Sources register ​

The evidence behind the skill's slow-rotting claims. Every claim the skill makes about the outside world that will not change week to week has an entry here; Mops answers "where did you get this?" — in any phrasing — from this file (REFERENCE §7, SKILL "Say what you know"). It is read by whichever flow needs it, and it is read first — a register entry serves every flow that meets its need, and the live web is where the register runs out, not where it starts (FLOWS → a shelf serves every flow).

What belongs here — and what never does. The register holds slow-rotting canon only: findings, methods and standards that age in years. Fast-rotting facts never enter — a price, a current API limit, a competitor's live feature stay fetch-at-decision-time rules, quoted with their check-date at the moment of use, never cached here to rot.

One fixed form per entry, so a wrong entry is visibly wrong: id · full citation · live URL/DOI/arXiv · archive link · licence · one-paragraph distillate (our words) · check-date · cited-by

Reads against is the field that stops a register becoming a pile. A register that can only answer "what do we have on X" will hold two entries pulling opposite ways and never say so. The field names the entries this one must be read beside and what the tension is, or one of two words that are not the same answer: none found — looked, nothing here pulls against it — or not checked, honest ignorance. An empty field is a defect; the two words are not. It is symmetric, and python3 scripts/fetch-source.py --verify-reads refuses the pair that only points one way. Ported from the sibling 2026-09-10, where it caught two one-way pairs on its first run; here it reported 15 of 15 still not checked on arrival, which was a fact about this register rather than a defect in it. The first pair to name each other is the tool-format pair at the end.

A back-pointer names a section and proves what it said — file.md#anchor (sha:…, checked …) — minted by scripts/fetch-source.py --cite <file.md>#<anchor>. A line number names a position, and a position moves the next time a paragraph is inserted above it: measured 2026-08-07, 11 of 23 line-number pointers no longer landed on their claim, with no edit to this register in between. The hash earns the rest — a passage rewritten underneath its citation makes the fact unknown and says so, where a line number cannot even see it. The one thing the anchored form cannot cite is a passage with no heading above it; that is a reason to give a load-bearing passage a heading, and until then the line form is used and marked as such.

Licence tiers (they decide what we may hold): free — an open licence (MIT, CC, a public standard); a copy may be carried later. copyrighted — citation + archive link + our own distillate, never the text itself. math — a formula recorded by name, which is not copyrightable. Each entry states its tier.

Upkeep. scripts/fetch-source.py builds and checks these entries: --resolve <doi|arxiv|url> prints a skeleton, --archive <url> triggers a Wayback snapshot, --verify walks every live URL. --verify-citations walks the other edge — every cited-by back into the doc that cites it — naming any pointer a rewrite left behind, and any entry the skill no longer cites at all (one parked for a later release stays named every run, which is the point: a deferral nobody is reminded of is a deletion). Both run each release (AGENTS.md → Cutting a release).

Persona theatre — grounding, fidelity, and the limits of synthetic audiences ​

park-self-reports · Park et al., self-report-grounded individual simulation ​

  • Citation: Park, J.S., et al. "LLM Agents Grounded in Self-Reports Enable General-Purpose Simulation of Individuals." arXiv:2411.10109 (2024; v1 was "Generative Agent Simulations of 1,000 People").
  • Live: https://arxiv.org/abs/2411.10109
  • Archive: http://web.archive.org/web/20260726212521/https://arxiv.org/abs/2411.10109
  • Licence: copyrighted (author © under arXiv's non-exclusive distribution licence) — cite + archive + our distillate
  • Distillate: Agents built from a person's own self-reports reproduce that person's survey answers at 83% (interview-grounded) / 82% (survey-grounded) / 86% (both) of the person's two-week test-retest ceiling, versus 74% for demographics-only; a free-text "persona paragraph" scores 0.71, below even the demographics baseline (0.74). Self-report grounding also reduces accuracy disparities across racial and ideological groups. The takeaway the skill leans on: the grounding artifact — the interview transcript — is the product, not a written bio.
  • Check-date: 2026-07-27
  • Reads against: not checked
  • Cited-by: MODULES.md#persona-theatre-synthetic-and-live-audiences (sha:070ca2bf, checked 2026-09-11), MODULES.md#staging-proto-persona-validated-persona (sha:f8ca5f73, checked 2026-08-07), templates/PERSONA-template.md#bias-profile-24-named-biases-each-with-its-source (sha:80d2a6dd, checked 2026-08-07)

park-hai-brief · Park et al., Stanford HAI policy brief ​

ashokkumar-nature · Ashokkumar et al., direction not magnitude ​

  • Citation: Ashokkumar, A., Hewitt, L., Ghezae, I., Willer, R. Nature (advance online publication, 2026-07-08). doi:10.1038/s41586-026-10742-x.
  • Live: https://doi.org/10.1038/s41586-026-10742-x
  • Archive: pending (Save Page Now triggered 2026-07-27; availability: https://archive.org/wayback/available?url=https://doi.org/10.1038/s41586-026-10742-x)
  • Licence: copyrighted (Springer Nature) — cite + archive + our distillate
  • Distillate: Across a large replication set, LLM simulations track the direction of experimental effects at about r≈0.85 while systematically overestimating their magnitude. This is the evidence for the theatre's hardest rule: a synthetic verdict may state direction, never a magnitude (no "23% would churn").
  • Check-date: 2026-07-27
  • Reads against: not checked
  • Cited-by: MODULES.md#persona-theatre-synthetic-and-live-audiences (sha:070ca2bf, checked 2026-09-11)

ls-types · Lewis & Sauro, a taxonomy of synthetic users ​

ls-review · Lewis & Sauro, a review of synthetic-user experiments ​

  • Citation: Lewis, J., Sauro, J. "A Review of Experiments with Synthetic Users." MeasuringU (2026-04-14).
  • Live: https://measuringu.com/review-of-experiments-with-synthetic-users/
  • Archive: http://web.archive.org/web/20260512065453/https://measuringu.com/review-of-experiments-with-synthetic-users/
  • Licence: copyrighted (MeasuringU) — cite + archive + our distillate
  • Distillate: Reviews ~12 recent experiments with synthetic users and finds mixed results, with synthetic responses showing artificially low variability and distorted magnitudes relative to real respondents — so they can indicate direction but not the size of an effect. (Their framing — low variability and distortion — is what the skill states, not "clustering toward neutral.")
  • Check-date: 2026-07-27
  • Reads against: not checked
  • Cited-by: MODULES.md#persona-theatre-synthetic-and-live-audiences (sha:070ca2bf, checked 2026-09-11), MODULES.md#accuracy-score-and-consent-for-twins-of-real-people (sha:59e1ec81, checked 2026-09-11)

sharma-sycophancy · Sharma et al., sycophancy is trained in ​

  • Citation: Sharma, M., et al. "Towards Understanding Sycophancy in Language Models." ICLR 2024. arXiv:2310.13548 (2023).
  • Live: https://arxiv.org/abs/2310.13548
  • Archive: http://web.archive.org/web/20260725125159/https://arxiv.org/abs/2310.13548
  • Licence: copyrighted (author © under arXiv's non-exclusive distribution licence) — cite + archive + our distillate
  • Distillate: Sycophancy — telling the user what they want to hear — is a trained-in property of RLHF'd assistants, consistent across several models and tasks. A persona built on such a model inherits that compliance, which is why the theatre's calibration layer suppresses sycophancy explicitly (a synthetic respondent is a pleaser unless corrected).
  • Check-date: 2026-07-27
  • Reads against: not checked
  • Cited-by: MODULES.md#bias-profiles-every-persona-carries-24-each-with-its-source (sha:a01bd41f, checked 2026-08-07)

tjuatja-biases · Tjuatja et al., LLM response biases ≠ human ones ​

  • Citation: Tjuatja, L., et al. "Do LLMs Exhibit Human-like Response Biases? A Case Study in Survey Design." TACL 12 (2024). arXiv:2311.04076 (2023).
  • Live: https://arxiv.org/abs/2311.04076
  • Archive: http://web.archive.org/web/20260116061015/https://arxiv.org/abs/2311.04076
  • Licence: copyrighted (author © under arXiv's non-exclusive distribution licence) — cite + archive + our distillate
  • Distillate: Tests whether LLMs reproduce known human survey response biases (acquiescence, question-order effects) and finds their biases do not reliably mirror human ones — sometimes absent, sometimes inverted. Caveats how far a synthetic survey respondent can stand in for a human one.
  • Check-date: 2026-07-27
  • Reads against: not checked
  • Cited-by: theatre canon (skill distillation pending 2.7)

argyle-silicon · Argyle et al., silicon sampling and its diversity limits ​

  • Citation: Argyle, L.P., et al. "Out of One, Many: Using Language Models to Simulate Human Samples." Political Analysis (2023). doi:10.1017/pan.2023.2.
  • Live: https://doi.org/10.1017/pan.2023.2
  • Archive: pending (Save Page Now blocked HTTP 523 on 2026-07-27; availability: https://archive.org/wayback/available?url=https://doi.org/10.1017/pan.2023.2)
  • Licence: copyrighted (Cambridge University Press) — cite + archive + our distillate
  • Distillate: Introduces "silicon sampling" — conditioning an LLM on demographic backstories to simulate human survey samples — and shows it can reproduce some subgroup patterns while collapsing within-group diversity. Backs the caution that synthetic samples flatten variety rather than represent it.
  • Check-date: 2026-07-27
  • Reads against: not checked
  • Cited-by: theatre canon (skill distillation pending 2.7)

wang-flattening · Wang et al., identity flattening ​

  • Citation: Wang, A., Morgenstern, J., Dickerson, J.P. "Large language models that replace human participants can harmfully misportray and flatten identity groups." Nature Machine Intelligence (2025). arXiv:2402.01908.
  • Live: https://arxiv.org/abs/2402.01908
  • Archive: http://web.archive.org/web/20260607174738/https://arxiv.org/abs/2402.01908
  • Licence: copyrighted (author © under arXiv's non-exclusive distribution licence; journal © Springer Nature) — cite + archive + our distillate
  • Distillate: Finds that using LLMs to replace human participants can harmfully misportray and flatten identity groups — reproducing majority stereotypes and erasing within-group variation. This is the direct evidence for the never-assign-a-bias-from-demographics rule: a demographic backstory produces a caricature, not a person.
  • Check-date: 2026-07-27
  • Reads against: not checked
  • Cited-by: MODULES.md#staging-proto-persona-validated-persona (sha:f8ca5f73, checked 2026-08-07), templates/PERSONA-template.md#bias-profile-24-named-biases-each-with-its-source (sha:80d2a6dd, checked 2026-08-07)

kapania-simulacrum · Kapania et al., LLMs as qualitative participants ​

  • Citation: Kapania, S., et al. "'Simulacrum of Stories': Examining Large Language Models as Qualitative Research Participants." CHI 2025. arXiv:2409.19430 (2024).
  • Live: https://arxiv.org/abs/2409.19430
  • Archive: http://web.archive.org/web/20260411143601/https://arxiv.org/abs/2409.19430
  • Licence: copyrighted (author © under arXiv's non-exclusive distribution licence) — cite + archive + our distillate
  • Distillate: Treating LLMs as qualitative research participants yields plausible but hollow "simulacra of stories" that miss the lived specificity of real interviews. Marks the boundary of synthetic personas in qualitative work — a supplement, never a replacement for a real transcript.
  • Check-date: 2026-07-27
  • Reads against: not checked
  • Cited-by: theatre canon (skill distillation pending 2.7)

Cost routing — cheap-first, conditional on a good verifier ​

frugalgpt · Chen, Zaharia & Zou, FrugalGPT ​

  • Citation: Chen, L., Zaharia, M., Zou, J. "FrugalGPT: How to Use Large Language Models While Reducing Cost and Improving Performance." arXiv:2305.05176 (2023).
  • Live: https://arxiv.org/abs/2305.05176
  • Archive: http://web.archive.org/web/20260722202256/https://arxiv.org/abs/2305.05176
  • Licence: copyrighted (author © under arXiv's non-exclusive distribution licence) — cite + archive + our distillate
  • Distillate: A cascade that queries cheaper models first and escalates only on low confidence can match the best single model's accuracy at up to −98% cost. The evidence for cheap-first-then-escalate routing at decomposition.
  • Check-date: 2026-07-27
  • Reads against: not checked
  • Cited-by: ROLES.md#grades-fit-check-and-the-talent-pool (sha:45943d7c, checked 2026-09-25)

routerbench · Hu et al., RouterBench ​

  • Citation: Hu, Q.J., et al. "RouterBench: A Benchmark for Multi-LLM Routing Systems." arXiv:2403.12031 (2024).
  • Live: https://arxiv.org/abs/2403.12031
  • Archive: http://web.archive.org/web/20260606022825/https://arxiv.org/abs/2403.12031
  • Licence: copyrighted (author © under arXiv's non-exclusive distribution licence) — cite + archive + our distillate
  • Distillate: Cascades beat both any individual LLM and a zero-cost router only when the verifier is good — judge error ≤0.1, deteriorating past 0.2. The load-bearing caveat: cheap-first routing is conditional on a good verifier. In the skill, the review gates are that verifier, so the condition is already met — the caveat reads as a strength, not a risk.
  • Check-date: 2026-07-27
  • Reads against: not checked
  • Cited-by: ROLES.md#grades-fit-check-and-the-talent-pool (sha:45943d7c, checked 2026-09-25)

Repository context files ​

agentsmd-eth · Gloaguen et al., do AGENTS.md files help? ​

  • Citation: Gloaguen, T., Mündler, N., Müller, M.N., Raychev, V., Vechev, M. (ETH Zurich). "Evaluating AGENTS.md: Are Repository-Level Context Files Helpful for Coding Agents?" arXiv:2602.11988 (v2, 2026-06-23).
  • Live: https://arxiv.org/abs/2602.11988
  • Archive: http://web.archive.org/web/20260711121106/https://arxiv.org/abs/2602.11988
  • Licence: copyrighted (author © under arXiv's non-exclusive distribution licence) — cite + archive + our distillate
  • Distillate: A coding-agent benchmark finds repository-level context files (AGENTS.md) do not improve task success rates and add roughly +20% inference cost; LLM-generated context files perform slightly worse than none. The evidence behind "curate the shared guide, don't autogenerate it."
  • Check-date: 2026-07-27
  • Reads against: not checked
  • Cited-by: skills/mops/SKILL.md:78 (line form: the claim sits in the core's preamble, above its first heading, so there is no section to anchor to)

Method provenance and standards (references, not evidence claims) ​

method-provenance · adapted methods, nothing embedded ​

  • Citation: cookiy — user-research-skill (MIT). agentman — "Synthetic Persona Creator" skill (concepts only).
  • Live: https://github.com/cookiy-ai/user-research-skill · https://agentman.ai/agentskills/skill/synthetic-persona-creator (owner-confirmed 2026-07-27; page verified live: three-layer persona architecture — identity foundation · context seeding · response calibration — plus cohort distribution)
  • Archive: http://web.archive.org/web/20260411081018/https://github.com/cookiy-ai/user-research-skill
  • Licence: free — cookiy's user-research-skill is MIT (a copy may be carried later; today only the method shape is adapted, nothing embedded). agentman: no licence stated on the page; concepts taken, nothing embedded (no code carried, so no licence obligation).
  • Distillate: Method lineage, not evidence. cookiy's MIT skill supplied the shape of the qualitative-research flows (its qualitative-research-planner → our persona-interview flow, its synthesize-research-report → our QDA step), adapted through the import gate. agentman supplied the calibration and cohort concepts behind the persona response-calibration layer. Recorded so every adaptation is auditable and no vendor wrapper is smuggled in.
  • Check-date: 2026-07-27
  • Reads against: not checked
  • Cited-by: MODULES.md#staging-proto-persona-validated-persona (sha:f8ca5f73, checked 2026-08-07), MODULES.md#bias-profiles-every-persona-carries-24-each-with-its-source (sha:a01bd41f, checked 2026-08-07), MODULES.md#mixed-live-synthetic-hypothesis-beside-fact (sha:5f160353, checked 2026-08-07), ROLES.md#any-role-from-conversation-the-role-builder (sha:76e2813c, checked 2026-09-11)

standards-cluster · named review standards ​

claim-audit · academic-research-skills — a citation that resolves is not one that supports ​

  • Citation: Imbad0202, academic-research-skills — its claim-level citation audit and the post-publication audit report in examples/showcase/ (README read 2026-09-25).
  • Live: https://github.com/Imbad0202/academic-research-skills
  • Archive: archive: pending (run: fetch-source.py --archive https://github.com/Imbad0202/academic-research-skills)
  • Licence: CC BY-NC 4.0 — non-commercial, so the method is taken and none of the text or code; this company runs commercial work
  • Distillate: Checking that a cited source exists says nothing about whether the sentence carrying it says what the source says. The pack reads each claim against the passage it cites and classes it as supported, overstated, unsupported, or unverifiable when the passage cannot be reached. Its authors report that an independent audit after publication found 21 of 68 real issues that three rounds of integrity checks had missed — the lesson of a review that reports clean, measured in another field. Not re-measured here.
  • Check-date: 2026-09-25
  • Reads against: none found
  • Cited-by: STACKS.md#evidence-sources-a-claim-can-actually-rest-on (sha:b15d7112, checked 2026-09-25)

What the tools print — read in their own sources ​

paper-lookup-200 · K-Dense, paper-lookup — scholarly APIs that answer a failure with HTTP 200 ​

  • Citation: K-Dense-AI, scientific-agent-skills, skills/paper-lookup/SKILL.md (main branch, read 2026-09-25).
  • Live: https://github.com/K-Dense-AI/scientific-agent-skills/blob/main/skills/paper-lookup/SKILL.md
  • Archive: archive: pending (run: fetch-source.py --archive https://github.com/K-Dense-AI/scientific-agent-skills/blob/main/skills/paper-lookup/SKILL.md)
  • Licence: MIT — a copy may be carried; only our distillate is held
  • Distillate: Six free scholarly APIs report a failure as success. PMC returns a well-formed article with no body when the publisher forbids redistribution; arXiv returns one entry titled Error for a malformed parameter and quietly widens an unknown field prefix to all fields; Europe PMC puts its error code inside a 200 body; bioRxiv accepts an out-of-step pagination cursor and returns the wrong thirty records; Figshare's GET search ignores its query; OpenCitations answers a DOI that does not exist with a count of zero. So the body is read before the answer is believed — and the zero is the dangerous one, being a plausible fact about a real paper. Not re-measured here.
  • Check-date: 2026-09-25
  • Reads against: none found
  • Cited-by: STACKS.md#evidence-sources-a-claim-can-actually-rest-on (sha:b15d7112, checked 2026-09-25)

skillspector-json · SkillSpector — a rejected manifest shows only in the JSON ​

  • Citation: NVIDIA, SkillSpector, the scan command's JSON report (--format json --output <file>), run on this repository's skills 2026-09-24.
  • Live: https://github.com/NVIDIA/skillspector
  • Archive: archive: pending (run: fetch-source.py --archive https://github.com/NVIDIA/skillspector)
  • Licence: Apache-2.0 — a copy may be carried; only our distillate is held
  • Distillate: A skill whose frontmatter a strict YAML parser rejects is skipped, and the printed summary still says the scan succeeded with no findings. The only trace is a manifest_parse_error in that skill's analysis_completeness.ledger_exceptions — in a --recursive report under skills[], while the report's own top-level analysis_completeness stays clean. Measured on this repository and its sibling's thirteen descriptions holding a plain value with : inside it; the repair reads the JSON, where grep -c manifest_parse_error must print 0.
  • Check-date: 2026-09-24
  • Reads against: none found
  • Cited-by: AGENTS.md#cutting-a-release (sha:1f6daa8d, checked 2026-09-25), SECURITY.md#third-party-skills (sha:643fb065, checked 2026-09-25)

mise-outdated · mise, outdated — what its JSON means, and what makes it refuse a file ​

  • Citation: jdx/mise, src/toolset/outdated_info.rs (main branch, read 2026-09-11), and the mise outdated, mise ls and mise bootstrap help of mise 2026.9.5.
  • Live: https://github.com/jdx/mise/blob/main/src/toolset/outdated_info.rs
  • Archive: archive: pending (run: fetch-source.py --archive https://github.com/jdx/mise/blob/main/src/toolset/outdated_info.rs)
  • Licence: MIT — a copy may be carried; only our distillate is held
  • Distillate: mise outdated --json is an object keyed by tool — requested, current (null when not installed), bump, latest, and release_url (omitted when the backend publishes none). latest means the newest version inside the pin without --bump, and the newest at all with it; bump is the new pin at the old pin's precision (22 becomes 24). A tool whose lookup fails is dropped from the JSON with only a warning on stderr, so an offline run and a current one both print {}. The short -l is deprecated and becomes --local in 2027.8.5. Measured on 2026.9.5: a mise.toml with [tools] alone is read on any machine, while one carrying [bootstrap.packages] is refused by every command until mise trust.
  • Check-date: 2026-09-11
  • Reads against: npm-outdated — they agree, and that is the finding: npm's wanted is mise's latest without --bump, npm's latest is mise's with it, so one report can sort both by the same line
  • Cited-by: PLAYBOOKS.md#when-something-it-needs-has-a-newer-version (sha:53fbb8b2, checked 2026-09-25), PLAYBOOKS.md#what-the-project-needs-from-the-machine (sha:0d4ca08d, checked 2026-09-25)

npm-outdated · npm, npm outdated — an update inside the range, and a move beyond it ​

  • Citation: npm CLI documentation, npm-outdated, and lib/commands/outdated.js, npm/cli latest branch, read 2026-09-11.
  • Live: https://docs.npmjs.com/cli/commands/npm-outdated
  • Archive: archive: pending (run: fetch-source.py --archive https://docs.npmjs.com/cli/commands/npm-outdated)
  • Licence: copyrighted (npm, a GitHub company) — cite + our distillate
  • Distillate: wanted is the newest version that satisfies the range in package.json; latest is the version tagged latest in the registry, and npm's own output colours the first update now and the second proceed with caution. With --json the answer is an object keyed by package name whose value becomes an array when one name is outdated in two places; current is absent when the package is not installed; the command exits 1 whenever anything is outdated, so its exit code says nothing about whether it answered.
  • Check-date: 2026-09-11
  • Reads against: mise-outdated — the same line drawn for runtimes and tools: they agree, which is what lets the pin, not the numbering, decide what is routine
  • Cited-by: PLAYBOOKS.md#when-something-it-needs-has-a-newer-version (sha:53fbb8b2, checked 2026-09-25)

pep-594 · Python 3.13 removed nineteen standard-library modules ​

  • Citation: Python Software Foundation, What's New In Python 3.13, "Important removals"; PEP 594, "Removing dead batteries from the standard library".
  • Live: https://docs.python.org/3/whatsnew/3.13.html
  • Archive: archive: pending (run: fetch-source.py --archive https://docs.python.org/3/whatsnew/3.13.html)
  • Licence: copyrighted (© Python Software Foundation) — cite + our distillate
  • Distillate: 3.13 removed the nineteen "dead batteries" deprecated in 3.11 — aifc, audioop, cgi, cgitb, chunk, crypt, imghdr, mailcap, msilib, nis, nntplib, ossaudiodev, pipes, sndhdr, spwd, sunau, telnetlib, uu and xdrlib — and the 2to3 tool with lib2to3. A minor version number that breaks any code importing one of them: the reason the line between a routine update and a decision is the project's own pin, not the version's numbering.
  • Check-date: 2026-09-11
  • Reads against: none found
  • Cited-by: PLAYBOOKS.md#when-something-it-needs-has-a-newer-version (sha:53fbb8b2, checked 2026-09-25)