farhoodlabs/cycling-training
8
0
Fork 0
Code Issues Pull Requests 1 Actions Packages Projects Releases 1 Wiki Activity

Add cycling-training Claude skill (Intervals.icu MCP companion) #1

Merged
Chris Farhood merged 3 commits from feat/cycling-training-skills into main 2026-07-20 23:13:23 +00:00
Conversation 0 Commits 3 Files Changed 13
+1023
Chris Farhood commented 2026-07-20 23:00:40 +00:00
Owner
Copy Link
Copy Source

Modular cycling-training Claude skill: periodization logic, workout design, and research-backed protocols that companion an Intervals.icu MCP server. Hard split — the MCP server owns all data + computed metrics (activities, wellness, power curves, CTL/ATL/TSB, zones, TSS); this skill owns knowledge + reasoning. No calculators that reproduce MCP math.

Structure

  • SKILL.md — router (routes to reference docs by need) + the data/knowledge split
  • README.md — companion-to-MCP architecture, install-as-a-skill instructions
  • references/ — durability, mtb-xco-demands, strength-for-cyclists, vo2max-intervals, periodization, data-confounds, citations
  • assets/ — base + build plan templates, the ~1,500 kJ to 1-min hill durability field test

Citations

Every quantitative claim carries an inline citation; references/citations.md is the single source of truth with PMID/DOI, each verified against PubMed on 2026-07-20 by independent verification agents. Confidence flags (confirmed / commonly-cited / not-found) are explicit.

Honesty flags baked in (verify, don't trust)

  • VO2 rep length is a genuine conflict: running (Fleckenstein 2025) favors long intervals; cycling (Ronnestad 2016/2020/2021) favors short 30/15s for time >=90% VO2max. Documented as a conflict — the cycling evidence cuts against a blanket "bias toward longer reps".
  • "Yu et al. 2025" (frequency > distribution) could not be located in PubMed — not cited; the frequency point is grounded in Seiler 2024 + Sun/Yu 2025 review and labelled a heuristic.
  • Under-researched doses flagged "track individual response": durability dose, taper TSB window, CTL ramp rate.
  • Data-confound doc forbids naive rules (no hardcoded decoupling<5%=fresh or RHR+5=overtrained); power governs, HR conditional, MTB HR/power/VO2 dissociation documented.
  • Discipline-specific numbers labelled; small-n elite XCO data not over-transferred to masters marathon.

🤖 Generated with Claude Code

https://claude.ai/code/session_01TqrhBhC3GEcKyaw8RTk8G6

Modular **cycling-training** Claude skill: periodization logic, workout design, and research-backed protocols that **companion an Intervals.icu MCP server**. Hard split — the MCP server owns all data + computed metrics (activities, wellness, power curves, CTL/ATL/TSB, zones, TSS); this skill owns knowledge + reasoning. No calculators that reproduce MCP math. ## Structure - `SKILL.md` — router (routes to reference docs by need) + the data/knowledge split - `README.md` — companion-to-MCP architecture, install-as-a-skill instructions - `references/` — durability, mtb-xco-demands, strength-for-cyclists, vo2max-intervals, periodization, data-confounds, citations - `assets/` — base + build plan templates, the ~1,500 kJ to 1-min hill durability field test ## Citations Every quantitative claim carries an inline citation; `references/citations.md` is the single source of truth with **PMID/DOI**, each verified against PubMed on 2026-07-20 by independent verification agents. Confidence flags (confirmed / commonly-cited / not-found) are explicit. ## Honesty flags baked in (verify, don't trust) - **VO2 rep length is a genuine conflict:** running (Fleckenstein 2025) favors long intervals; cycling (Ronnestad 2016/2020/2021) favors short 30/15s for time >=90% VO2max. Documented as a conflict — the cycling evidence cuts against a blanket "bias toward longer reps". - **"Yu et al. 2025" (frequency > distribution) could not be located in PubMed** — not cited; the frequency point is grounded in Seiler 2024 + Sun/Yu 2025 review and labelled a heuristic. - **Under-researched doses flagged** "track individual response": durability dose, taper TSB window, CTL ramp rate. - **Data-confound doc forbids naive rules** (no hardcoded decoupling<5%=fresh or RHR+5=overtrained); power governs, HR conditional, MTB HR/power/VO2 dissociation documented. - **Discipline-specific numbers labelled**; small-n elite XCO data not over-transferred to masters marathon. 🤖 Generated with [Claude Code](https://claude.com/claude-code) https://claude.ai/code/session_01TqrhBhC3GEcKyaw8RTk8G6
Chris Farhood added 1 commit 2026-07-20 23:00:41 +00:00
feat: add cycling-training Claude skill (MCP companion) f6dd9dfc27 
Modular cycling-training skill: periodization logic, workout design, and
research-backed protocols that companion an Intervals.icu MCP server (the
server owns all data + computed metrics; the skill owns knowledge + reasoning).

Structure:
- SKILL.md router + README (companion-to-MCP architecture, install steps)
- references/: durability, mtb-xco-demands, strength-for-cyclists,
  vo2max-intervals, periodization, data-confounds, citations
- assets/: base + build plan templates, durability field-test protocol

Every quantitative claim carries an inline citation; citations.md is the
single source of truth with PMID/DOI, each verified against PubMed on
2026-07-20. Notable honesty flags baked in:
- VO2 rep-length: running (Fleckenstein 2025) vs cycling (Ronnestad) conflict
  documented rather than asserted; cycling evidence favors short intervals.
- 'Yu et al. 2025' (frequency > distribution) could not be found; not cited.
- Under-researched doses (durability, taper TSB, CTL ramp) flagged 'track response'.
- HR-confound doc forbids naive decoupling<5% / RHR+5 rules.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqrhBhC3GEcKyaw8RTk8G6
Chris Farhood added 1 commit 2026-07-20 23:05:16 +00:00
fix(vo2max): rebalance rep-length conflict, correct overstatements 85b6be4e36 
Review of the VO2 conflict section against primary sources (verified via
PubMed + search):

- Rønnestad & Hansen 2016 (PMID 23942167) confirmed: 13 cyclists, 30s vs
  50%/80% Tmax at 2:1 work:rest, measured time >=90% VO2peak, 30s won. Table
  row corrected to reflect the actual comparison.
- Yang 2025 optimum ~140s is a MEDIUM rep (~2.3 min), between the poles.
  Section reframed from binary short-vs-long to an inverted-U with a
  modality/protocol-dependent peak. 'Weighted toward running' softened to
  match what the abstract actually supports.
- Removed the overstatement 'best-controlled evidence favors short 30/15
  intervals'. Added two caveats: (1) cycling short-interval superiority is
  largely one research group; (2) effort-matched work (total-work-matched)
  reportedly nulls the advantage (Seiler group; unpublished/secondary,
  flagged low-tier, not cited as fact).
- Practical menu now defaults to medium (~2-4 min, near the pooled optimum)
  rather than short; short/long presented as goal-dependent options.
- citations.md: added single-group caveat on Ronnestad body of work and an
  explicit NOT-VERIFIABLE entry for the effort-matched replication.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqrhBhC3GEcKyaw8RTk8G6
Chris Farhood added 1 commit 2026-07-20 23:11:41 +00:00
fix: resolve code-review correctness bugs in README + build template 4e4f719670 
Three CONFIRMED defects, all from template/README text written before the
vo2max conflict section was revised, now reconciled with the corrected docs:

- README taper example said TSB '−5 to +5' (under-recovered) — corrected to
  the doctrinal +5 to +15 that periodization.md and both plan templates use.
- plan-template-build.md VO2 prescription said 'bias toward longer reps
  (3–5 min)... accumulate more time >90% VO2max than 30/30s' — contradicted
  its own cited source and was self-contradictory (140s ≠ 3–5 min). Replaced
  with the neutral menu: default medium (~2–4 min, near the ~140s optimum),
  short/long by goal, 'no rep length proven superior in cycling.'
- plan-template-build.md cited the phantom 'Yu et al.' (marked ❌ NOT FOUND in
  citations.md) — re-grounded the frequency heuristic in Seiler 2024.

DRY findings (inline PMID/DOI duplication, restated rhythm/field-test tables)
left as-is: inline citations and self-contained, independently-loadable docs
were explicit design requirements for this skill.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TqrhBhC3GEcKyaw8RTk8G6
Chris Farhood merged commit 517986781f into main 2026-07-20 23:13:23 +00:00
Sign in to join this conversation.
Reviewers
No Reviewers
Labels
No items
No Label
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
fl_blake (Blake MIddleware) cpfarhood (Chris Farhood) ci (Continuous Integration [bot]) flux (Flux CD) admin (Gitea Admin) fl_hugh (Hugh Commit) renovate (Mend Renovate) fl_warren (Warren Bufferpool)
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: farhoodlabs/cycling-training#1
Delete Branch "feat/cycling-training-skills"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?