Bad
stephenleo/bmad-autonomous-development
BMad Autonomous Development — orchestrates parallel story implementation pipelines.
Executes an implementation plan produced by dev-plan in the role of a staff-level Engineer.
$ npx skills add FHIR/fhir-codegen --skill dev-do -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install FHIR/fhir-codegen dev-do --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
$ git clone --depth 1 https://github.com/FHIR/fhir-codegen.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.github/skills/dev-do .claude/skills/dev-do && rm -rf skills-srcUse ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.
Claude Code skills documentation · loads skills from .claude/skills/
Install the "dev-do" agent skill from https://github.com/FHIR/fhir-codegen/tree/main/.github/skills/dev-do into .claude/skills/dev-do/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "dev-do", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$skill-installer install https://github.com/FHIR/fhir-codegen/tree/main/.github/skills/dev-doType this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.
$ npx skills add FHIR/fhir-codegen --skill dev-do -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install FHIR/fhir-codegen dev-do --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/FHIR/fhir-codegen.git skills-src && mkdir -p .agents/skills && cp -r skills-src/.github/skills/dev-do .agents/skills/dev-do && rm -rf skills-srcUse ~/.agents/skills/ instead of .agents/skills for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "dev-do" agent skill from https://github.com/FHIR/fhir-codegen/tree/main/.github/skills/dev-do into .agents/skills/dev-do/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "dev-do", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add FHIR/fhir-codegen --skill dev-do -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install FHIR/fhir-codegen dev-do --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/FHIR/fhir-codegen.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/.github/skills/dev-do .cursor/skills/dev-do && rm -rf skills-srcUse ~/.cursor/skills/ instead of .cursor/skills for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "dev-do" agent skill from https://github.com/FHIR/fhir-codegen/tree/main/.github/skills/dev-do into .cursor/skills/dev-do/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "dev-do", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gemini skills install https://github.com/FHIR/fhir-codegen.git --path .github/skills/dev-do--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.
$ npx skills add FHIR/fhir-codegen --skill dev-do -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install FHIR/fhir-codegen dev-do --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/FHIR/fhir-codegen.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/.github/skills/dev-do .gemini/skills/dev-do && rm -rf skills-srcUse ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "dev-do" agent skill from https://github.com/FHIR/fhir-codegen/tree/main/.github/skills/dev-do into .gemini/skills/dev-do/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "dev-do", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install FHIR/fhir-codegen dev-doInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add FHIR/fhir-codegen --skill dev-do -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
$ git clone --depth 1 https://github.com/FHIR/fhir-codegen.git skills-src && mkdir -p .github/skills && cp -r skills-src/.github/skills/dev-do .github/skills/dev-do && rm -rf skills-srcUse ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "dev-do" agent skill from https://github.com/FHIR/fhir-codegen/tree/main/.github/skills/dev-do into .github/skills/dev-do/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "dev-do", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add FHIR/fhir-codegen --skill dev-do -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install FHIR/fhir-codegen dev-do --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/FHIR/fhir-codegen.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/.github/skills/dev-do .opencode/skills/dev-do && rm -rf skills-srcUse ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "dev-do" agent skill from https://github.com/FHIR/fhir-codegen/tree/main/.github/skills/dev-do into .opencode/skills/dev-do/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "dev-do", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
dev-doExecutes an implementation plan produced by dev-plan in the role of a staff-level Engineer.
Dev Do is an agent skill from FHIR/fhir-codegen. Executes an implementation plan produced by dev-plan in the role of a staff-level Engineer. USE FOR: actually doing the work — writing/modifying code, running builds and tests, committing locally as phases complete, and keeping plan.md updated with current status. Accepts either a full path to the plan file or a short slot number that expands to scratch/[MMDD]-[]/plan.md. Optional maxsubagents (default 3) caps parallel sub-agent fan-out; optional checkpointevery (default 0 = never) yields back to the user after…
Its SKILL.md is about 5.3k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.
It sits in Agent Workflows, covering Subagents, Pull requests and Planning. It works with GitHub. The repository describes itself as: Tools for code generation based on the FHIR specification. The licence is MIT.
3 steps, taken from the first numbered list in SKILL.md.
Read from SKILL.md and the folder at commit b5f97c7. It shows what the files ask for, not the result of running them.
Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.
From allowed-tools in the SKILL.md frontmatter.
Shell commands in SKILL.md call:
gitghFrom the folder's file list and the shell code blocks in SKILL.md.
No URLs in SKILL.md. Its commands use git and gh, which can reach the network depending on how they are called.
From URLs in SKILL.md, links to its own repository left out.
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
Dev Do loads about 5.3k tokens when it runs. Until then it costs about 247 tokens; SKILL.md has 2,875 words of instructions outside code blocks.
Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.
The automated check found no risky patterns in SKILL.md.
Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.
The full file from FHIR/fhir-codegen at commit b5f97c7, republished under its MIT licence (© FHIR). 2,875 words, ~5,259 tokens.
.claude/skills/dev-do/SKILL.md (or your agent's skills folder).Acts as a staff-level Engineer for local development work in this
repository. Reads a plan.md (produced by dev-plan), implements it
phase by phase, runs builds/tests, and commits locally as it goes.
This skill is for shortcutting the local inner loop: it operates against the user's current working tree and may produce real commits. It does not push, and it does not open pull requests.
You are a staff-level Engineer. That means:
plan.md, and keep
going.max_subagents). Trivial single-file edits stay with you.Source (required) — where to read the plan. One of:
plan.md. Used
verbatim. Example: scratch/0423-02/plan.md.2, 02, 14).
Expands to scratch/<MMDD>-<##>/plan.md, where:<MMDD> is today's local date (zero-padded month + day).<##> is the slot number, always zero-padded to two digits.plan.md does not exist, stop and tell the user;
do not create one (that's dev-plan's job).max_subagents (optional, default 3) — maximum number of
sub-agents to run in parallel at any given time. 1 disables
parallel fan-out entirely. Hard upper bound: 8.
checkpoint_every (optional, default 0) — non-negative
integer. When 0 (the default), the skill runs all remaining
Pending phases back-to-back without ever pausing for user input.
When > 0, after every N successfully Complete phases the
skill posts a brief progress summary and yields so the user can
review or course-correct before the next phase starts. Blocked
phases, scope-exceeded decisions, pre-flight inconsistencies, and
final completion are separate yield conditions and always fire
regardless of this setting (see "Yield Conditions" below).
This skill is designed to drive a plan.md to completion in a single
invocation. The default contract is:
Pending phases back-to-back, while safely
reconciling any recorded In-progress or Blocked phase first.Explicit anti-patterns — do not do these:
dev-do between ordinary phases.
Re-invocation is the recovery path, except for the explicit
self-modification reload requirement below.These are the only reasons to stop and hand control back to the user mid-plan:
Blocked after reasonable debugging effort. Mark it
Blocked in plan.md with a one-line reason and stop.Complete — proceed to final verification, then
"Final Wrap-up" only if that gate succeeds.checkpoint_every > 0 and N phases have been marked Complete
since the last checkpoint. Post a brief progress summary
(commits + remaining phases) and yield.plan.md claims a phase is Complete but the working tree
disagrees.dev-do skill,
whether or not that change was committable. Record the durable
result and stop so a fresh invocation can reload the new
instructions before another phase begins.Anything else — including the satisfying click of a green test run —
is not a yield condition. Continue immediately to the next
Pending phase.
plan.md is the source of truth for what has been done. It is a control
file, not a work product: it lives in the gitignored slot directory, it is
never an owned path, and it is never staged, committed, or subjected to the
owned-path cleanliness checks. Editing it is always in scope.
plan.md is the sole exception to the ownership rules. Every other
repository path you touch — source, tests, documentation, configuration,
project files — must be declared under the phase's **Owned paths:**
before you edit it.
You must:
**Status:** line as you progress
(Pending → In-progress → Complete, or Blocked with a one-line
reason).Ready-to-execute → In-progress → Complete, or Blocked with an
actionable reason.## Progress Log section if not present and append entries in
the canonical PENDING / COMMIT / NOTE forms that dev-plan
defines. Before committing, append the PENDING entry carrying the
pre-commit HEAD, the staged tree ID, and the exact changed-path
list. Replace that entry with a COMMIT entry only after post-commit
identity checks pass.**Deviation:** sub-bullet.You must not delete plan.md and you must not delete the
sibling source request (featurerequest.md / bugreport.md).
plan.md and the sibling
source request (read-only) for context. Reading plan.md includes
reading its Issue row, which decides whether phase commits carry an
Issue: #N trailer.Draft — stop. The plan is not finished. Tell the user to
complete it with dev-plan first; do not implement it.Ready-to-execute — normal start; begin at the first Pending
phase.In-progress / Blocked — recovery start; apply the "Iteration
Mode" rules before touching anything.Complete — if every phase is also Complete, there is nothing
to do; report it is already done. If any phase is not Complete,
the plan is structurally corrupt: stop and report the
inconsistency rather than re-running a final gate that cannot
legitimately pass.AGENTS.md at the repository root — it is the canonical source for
build, test, and lint commands, code style, architectural
invariants, and commit trailers. If it is absent, fall back to
README.md / CONTRIBUTING.md and state in your output which
source you used. Never invent a command. Confirm every phase
verification command and every command under ## Final Verification
is sanctioned by AGENTS.md and has an unambiguous scope. If the
plan names a command AGENTS.md does not sanction, stop and ask
rather than guessing a substitute. For a legacy plan without a final
gate, pick a repository-valid build/test command from AGENTS.md
before editing anything; stop for clarification if the correct scope
is ambiguous.git diff --cached --quiet.plan.md, change
source, stage, unstage, reset, stash, or commit. Report that the
user must commit, unstage, or otherwise resolve the staged work.git status --short --branch and note unrelated changes
without modifying them.Complete phase's **Owned paths:**. Every
entry must be a literal repository-relative path. Before a
Pending phase starts, require every owned path to be completely
clean, including tracked, untracked, staged, and unstaged state.
Do not accept a pre-existing edit merely because it looks
compatible with the plan.git check-ignore. A
phase that owns a git-ignored path can never produce the commit
evidence a Complete status requires, so treat it as a plan
defect: mark the phase and plan Blocked and ask, rather than
force-adding the path or completing the phase without a commit.In-progress or Blocked phase, use the recovery rules
below. Do not simply rerun its steps.Complete lacks matching durable commit
evidence, stop on the inconsistency.Complete but the top-level status is
In-progress or Blocked, skip phase work and rerun final
verification.while loop, not a single pass.
Before every Pending phase, repeat the clean-index gate and owned
path cleanliness check. Then:Pending to
make ownership exhaustive. If another required path is found,
add it under that phase's **Owned paths:**, verify it is clean,
and only then continue. Never edit an undeclared path.In-progress.max_subagents.Verification command. If a command fails,
debug within reasonable scope. If it cannot be made green, mark
the phase and plan Blocked with an actionable reason and stop.git diff --cached --quiet again. A non-empty index is a
scope failure: leave it untouched, mark the phase and plan
Blocked, and stop.git diff --cached --name-only and the complete staged patch.
Require every staged path to belong to the phase and every
intended phase change to be present. On mismatch, do not unstage
or rewrite anything; mark Blocked and stop.HEAD, the exact staged changed-path list,
and the staged tree from git write-tree. Append the canonical
PENDING Progress Log entry carrying all three values. Keep the
phase In-progress.git commit --only -- <owned-paths>.
This path-limited form is mandatory even after staged-scope
inspection. Include every commit trailer required by AGENTS.md.
When the plan's Issue row names #N, append an Issue: #N
trailer alongside them. When that row says not published, or is
absent entirely, add nothing — an unbound slot produces exactly
the message it produced before this trailer existed.HEAD, its tree equals the recorded staged
tree, and its exact changed-path set equals the recorded list.
If commit creation or any identity check fails, do not amend,
reset, or otherwise rewrite the commit/index. Mark the phase and
plan Blocked with the evidence and stop.PENDING Progress Log entry with a COMMIT entry
carrying the actual SHA and subject, then mark the phase
Complete. Only this post-commit update may claim completion.checkpoint_every and
proceed to the next Pending phase.Complete, keep the
top-level plan In-progress and run every command under
## Final Verification. If any command remains red after reasonable
debugging, set the plan to Blocked with the failing command and
stop. Set the plan to Complete only after every command succeeds.
If a sanctioned verification cannot be run in this environment (for
example a gate needing setup AGENTS.md documents as a
prerequisite), do not claim it green — say explicitly which
verification you could not run and why.Complete, a Blocked phase, a scope-exceeded decision,
or a checkpoint boundary). Never fires per-phase. Report:dev-review against the same slot before
opening a PR, when the change is non-trivial — and then
dev-pr-open against the same slot when the user is ready to push
and open it.max_subagents cap is a concurrency cap, not a total cap.
You may launch more than max_subagents sub-agents over the life of
the task as long as no more than max_subagents are running at the
same time.code-review for an existing diff. When an
adversarial critique is useful and no registered specialist exists,
use a general-purpose sub-agent explicitly prompted for that role.AGENTS.md's commit conventions — it owns the sanctioned
type list, scope usage, subject style and length, and the required
trailers. Read it rather than assuming; repositories differ. Where
the session runtime supplies trailers of its own
(session/correlation ids), include those too.plan.md before executing.git commit --only -- <owned-paths> after staged-tree capture.Issue: #N trailer only when the plan's Issue row names
#N. It sits alongside the trailers AGENTS.md requires, and it
has no effect on the post-commit identity checks, which compare
parent, tree, and changed paths only — never the message.git push. Never gh pr create. Never force-push, amend,
or rewrite history as automatic recovery. Local phase commits only.
Pushing and opening a PR belong to dev-pr-open.A single dev-do invocation is expected to drive plan.md to
completion in one shot. Re-invocation is the recovery path — used
after a Blocked phase, a scope-exceeded yield, an explicit
checkpoint_every boundary, or an interrupted run — not the
normal mode of operation.
Before any recovery action, require git diff --cached --quiet. If the
index is non-empty, stop without changing it or plan.md.
When plan.md already shows In-progress, Blocked, or partial
completion:
Complete phase, its owned paths, status,
Progress Log, current HEAD, and working-tree state before deciding
what gate can safely resume.PENDING entry is durable recovery evidence only when it
contains the recorded base HEAD, staged tree ID, and exact changed
paths. If a candidate commit exists, reconcile it as the phase commit
only when its parent, tree, and changed paths exactly match all three
recorded values. Then replace it with a COMMIT entry and mark the
phase Complete.Blocked reason before retrying. Resume only when the
blocker is demonstrably resolved; otherwise report it unchanged.Complete disagrees with its commit evidence, stop
and ask the user rather than silently reconciling or redoing it.Complete while the plan is In-progress or
Blocked, rerun ## Final Verification; do not redo a phase or
report completion without that gate.dev-plan revision before implementation.<MMDD> for a numeric slot. For an earlier slot, the user
must give a full path.plan.md is editable, never deletable. Same for the sibling
source request.dev-plan.dev-pr-open is the skill that
pushes the branch and opens the pull request, and it exists precisely
so this rule never has to bend.Issue: #N trailer is conditional. It is added when — and
only when — the plan's Issue row names #N. The row itself belongs
to dev-issue; this skill reads it and never writes it.AGENTS.md. Read it before naming any build,
test, or lint command, and follow its code-style rules and
architectural invariants. If it is absent, fall back to README.md /
CONTRIBUTING.md and state which source you used. Never invent a
build or test command. Prefer the smallest scoped verification that
covers the change; escalate only when the scoped run says you must.
Where AGENTS.md is silent, match the surrounding code rather than
importing a preference from another repository. If a convention
contradicts the plan, prefer the convention and record the deviation
in plan.md.dev-do invocation that has loaded the new
instructions.Blocked, record why,
and report back.max_subagents sub-agents in parallel.© FHIR, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file
Just SKILL.md in .github/skills/dev-do of FHIR/fhir-codegen.
Open the folder on GitHubat commit b5f97c7
Dev Do next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| Dev Do this skillFHIR/fhir-codegen | 154 | — | ~5.3k | Automated safety check: Pass | MIT | |
| Badstephenleo/bmad-autonomous-development | 107 | — | ~7.7k | Automated safety check: Pass | MIT | |
| Implement FeatureDevBetterCom/DevBetterWeb | 157 | — | ~1.5k | Automated safety check: Pass | None | |
| Load PR CommentsNeoLabHQ/context-engineering-kit | 1.7k | — | ~2.1k | Automated safety check: Pass | GPL-3.0 | |
| Gh Issuestrpc-group/trpc-agent-go | 1.8k | 8 repos | ~8.7k | Automated safety check: Pass | Apache-2.0 | |
| GitHub Review Iterationprisma/orm | 48k | — | ~2.2k | Automated safety check: Pass | Apache-2.0 |
stephenleo/bmad-autonomous-development
BMad Autonomous Development — orchestrates parallel story implementation pipelines.
DevBetterCom/DevBetterWeb
End-to-end workflow for implementing, fixing, or otherwise working on a specific GitHub issue.
NeoLabHQ/context-engineering-kit
A skill your agent uses to load open/unresolved PR review comments then aggregate them as tasks in .specs/comments/.md for parallel agents to fix.
trpc-group/trpc-agent-go
Fetch GitHub issues, spawn sub-agents to implement fixes and open PRs, then monitor and address PR review comments.
prisma/orm
Runs a loop on a GitHub pull request: fetch review state, triage comments into actions, implement them and resolve threads, repeating until nothing actionable is left.
CherryHQ/cherry-studio
Reviews Cherry Studio branches, pull requests, commits, files and docs against the project's own architecture, naming, API-boundary and UI rules, report-only by default.
FHIR/fhir-codegen
Publishes a slot's feature request or bug report to GitHub as an issue, and keeps that issue in sync, in the role of a release-minded engineer.
FHIR/fhir-codegen
Drafts and iterates on local-development bug reports in the role of a staff-level Tech Lead.
FHIR/fhir-codegen
Drafts and iterates on local-development feature requests in the role of a staff-level Product Manager.
FHIR/fhir-codegen
Performs a two-track code-quality and QA review in the roles of a staff-level Engineering Lead and QA Lead, then synthesizes both critiques into a single analysis.md.
FHIR/fhir-codegen
Explores three competing solution shapes for one request in the roles of three isolated staff-level Engineering Leads, then has a fourth skeptical judge sub-agent select one on the record.
FHIR/fhir-codegen
Drives the entire local inner loop in one invocation, as a conductor over the skills that own each role.
Works with
Categories
Executes an implementation plan produced by dev-plan in the role of a staff-level Engineer. Dev Do is an agent skill from FHIR/fhir-codegen. Executes an implementation plan produced by dev-plan in the role of a staff-level Engineer.
Dev Do fits situations like: : actually doing the work — writing/modifying code; running builds and tests; committing locally as phases complete; keeping plan.md updated with current status.
Run `npx skills add FHIR/fhir-codegen --skill dev-do -a claude-code`. Or copy the skill folder (.github/skills/dev-do in FHIR/fhir-codegen) into .claude/skills/dev-do in your project. Claude Code loads it when a task matches its description.
Run `npx skills add FHIR/fhir-codegen --skill dev-do -a codex`. Or copy the skill folder (.github/skills/dev-do in FHIR/fhir-codegen) into .agents/skills/dev-do in your project. Codex loads it when a task matches its description.
Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add FHIR/fhir-codegen --skill dev-do -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/dev-do, .gemini/skills/dev-do, .github/skills/dev-do and .opencode/skills/dev-do in your project.
Going by SKILL.md and its folder, Dev Do needs the command-line tools its instructions call (git and gh).
SKILL.md contains no URLs. Its commands use git and gh, which can reach the network depending on how they are called. This is read from the text; nothing was executed.
Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.
Dev Do is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.
About 5.3k tokens (SKILL.md is roughly 21k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.
Skills that share tags, products or a category with Dev Do: Bad (stephenleo/bmad-autonomous-development, 107 stars), Implement Feature (DevBetterCom/DevBetterWeb, 157 stars), Load PR Comments (NeoLabHQ/context-engineering-kit, 1.7k stars) and Gh Issues (trpc-group/trpc-agent-go, 1.8k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
FHIR (a GitHub organization) maintains it in FHIR/fhir-codegen, which has 154 GitHub stars. The repository holds 9 skills in this directory. The repository was last updated on October 7, 2026.
Source: FHIR/fhir-codegen on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.