Skip to content

Add uipath-build-from-genome in its final form (1/6) [PILOT-9071] - #4327

Merged
RaduAna-Maria merged 4 commits into
mainfrom
feat/genome-stack-1-node-scripts
Nov 12, 2026
Merged

RaduAna-Maria merged 4 commits into
mainfrom
feat/genome-stack-1-node-scripts

Conversation

@RaduAna-Maria

@RaduAna-Maria RaduAna-Maria commented Nov 2, 2026 •

Copy link
Copy Markdown
Collaborator

Issue

PILOT-9071

Stack 1 of 6 (#4327 → #4328 → #4329 → #4330 → #4358 → #4379), merged together.

Summary

  • Add uipath-build-from-genome (build · check): build-side guides moved from uipath-genome (renames)
  • Replace check-genome.py with scripts/check-genome.mjs (Node, built-ins only)
  • Switch registries, README, CODEOWNERS and activation config from uipath-genome to the new skill
  • Name picked by the activation gate: 100% vs 85% for uipath-genome (run 41434186934)
  • check-genome.mjs context returns designed_in_fingerprint / designed_in_current; a changed SDD sends the genome to the planner's edit, then the build continues
  • An SDD's Execution autonomy no longer waives build questions; only the build request's own "take the defaults" does
  • A Decided: split and the Requested change topics are answered at build, never re-asked or confirmed

Paired: UiPath/migration-source-frameworks#110 moves the framework packs to uipath-planner / uipath-build-from-genome and check-genome.mjs; merge together.

Breaking: /uipath-genome is removed (rest in #4328).

🤖 Generated with Claude Code

@RaduAna-Maria
RaduAna-Maria added this pull request to stack #4331 November 2, 2026 10:53
@RaduAna-Maria
RaduAna-Maria force-pushed the feat/genome-stack-1-node-scripts branch from d33a4b3 to 3e14a65 Compare November 2, 2026 11:02
@RaduAna-Maria RaduAna-Maria changed the title Replace the genome checker and docx scripts with Node (1/4) [PILOT-9071] Replace the genome checker and docx scripts with Node (1/6) [PILOT-9071] Nov 3, 2026
@RaduAna-Maria
RaduAna-Maria force-pushed the feat/genome-stack-1-node-scripts branch from 3e14a65 to 9663f53 Compare November 4, 2026 13:34
@RaduAna-Maria RaduAna-Maria changed the title Replace the genome checker and docx scripts with Node (1/6) [PILOT-9071] Add uipath-build-from-genome in its final form (1/6) [PILOT-9071] Nov 4, 2026
@github-actions

github-actions Bot commented Nov 4, 2026 •

Copy link
Copy Markdown
Contributor

skill-lint: 0 new errors, 3 new warnings

Rule Location Message
warning LINK-BARE-PATH skills/uipath-genome/CLAUDE.md:15 mentions 'scripts/check-genome.py', which does not exist in the skill
warning LINK-BARE-PATH skills/uipath-genome/CLAUDE.md:24 mentions 'references/parts-guide.md', which does not exist in the skill
warning AGENT-READ-TRUNCATED skills/uipath-genome/references/transactional-execution-guide.md:1 40,068 bytes, 99 lines: Codex keeps 10,000 tokens (40,000 bytes) of one read. Split it or route to smaller files
What these rules check
  • AGENT-READ-TRUNCATED: Markdown files an agent reads should fit in one read of each target agent. Each agent caps what one read returns. Codex keeps 10,000 tokens, counted as UTF-8 bytes / 4, and drops the middle. Delegate's ReadFile, which LoadSkill also uses, returns the first 2,000 lines and at most 100,000 characters of them. Claude Code's Read returns the first ~25,000 tokens with a partial-view notice and, by default, refuses a whole file over 256 KB. Gemini CLI's read_file returns the first 2,000 lines and cuts lines over 2,000 characters. Past a cap the agent sees part of the file unless it reads it in pages. Split the file, or move detail into smaller references the skill routes to. SKILL.md is checked for Codex and Delegate only: Claude Code and Gemini load it whole when the skill activates. Claude Code's count is real tokens, estimated with claude_code_bytes_per_token; the others are exact.
  • LINK-BARE-PATH: File paths mentioned in text (not as links) must exist in the skill. Skills often tell the agent to 'read references/x.md' or 'run scripts/y.py' without a Markdown link, so link checkers never see them. A wrong path fails at the moment the agent needs it. Paths on a line that names another skill are resolved in that skill too.

Only what this PR adds is listed, compared with the commit it branched from; issues already on the base branch are not. Updated on every push.

@RaduAna-Maria
RaduAna-Maria force-pushed the feat/genome-stack-1-node-scripts branch 2 times, most recently from ad07ac7 to 6f5f204 Compare November 5, 2026 10:39

@danmunteanu1 danmunteanu1 left a comment

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

🟡 Commented · 5 blockers · 8 non-blockers
Inline findings on the files this PR changes; the stack-wide summary and the main-drift list are on #4379.

@danmunteanu1 danmunteanu1 left a comment

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

🟡 Commented · 6 blockers · 5 non-blockers
Re-review: new findings on this PR's fix commits; replies are on the earlier threads, and the stack-wide summary is on #4379.


The `context` call of [SKILL.md § Setup](../SKILL.md) has run; its fields decide the rows of § Read Plan below. `extracted` is true when the genome has a `## Source Map` that is not the stub; `transactional` is true when its Transactional Shape is not the stub, false for the stub, and null when the section is absent — null opens no transactional read and is a report item like the other findings below. `build_mode` `unknown` opens neither the migration row nor the refactor row until the Source Map's `Source framework` row, or the user's answer, names the framework ([SKILL.md § Setup](../SKILL.md)).

**A Build With or Components skill the host lacks** (found when the start row of § Read Plan invokes each named skill): name the skill and the components that need it, say the build needs it installed, and stop.

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

🔴 Blocker · A retired skill name now stops the build

A genome whose Build With names uipath-coded-workflows (a retired name) names a skill the host lacks, so this start-phase rule stops the build. The rule that a retired name "builds through uipath-rpa" (genome-reading-guide.md:33) is read only in the build phase (read-plan row :54), after the stop, and the checker reports a retired name as a report item, not a blocking FAIL.

Fix: except the retired names here ("…the host lacks, other than a retired name (genome-reading-guide.md § Build With and Components)"), and move the :54 row to the start phase.

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The build no longer stops on a retired name (e9efd78c7). The rule it now reaches (genome-reading-guide.md:33) builds all three retired names through uipath-rpa. But uipath-coded-agents was the Python coded agents skill, now uipath-agents (main e73afc26b), so a genome naming it gets an RPA project for an agent.

Fix: uipath-rpa-workflows and uipath-coded-workflows build through uipath-rpa; uipath-coded-agents builds through uipath-agents.

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Reproduced at bcf20b12a. I took the PR's exception-triage-agent-genome.md example and made its Build With row read `uipath-coded-agents` | Classifies the discrepancy and proposes a resolution. Then I asked for an unattended build, 3 runs (Opus 5.5, stopped at the first building skill). One run invoked uipath-rpa with "Build With row names retired skill uipath-coded-agents → builds through uipath-rpa. Coded C# workflow project"; the other two invoked uipath-agents. When the row's rationale says "Python coded agent", the agent overrides the rule (1/1). So the rule misroutes exactly the genomes that give no other hint.

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Downgrading this to a non-blocker. uipath-coded-agents was retired on 2026-04-02 and the planner's checker fails any genome that names it, so only a genome written before then can reach this rule. Simplest fix: drop uipath-coded-agents from the list here, so such a genome stops with the missing-skill rule (as main does) instead of building as an RPA project.

@RaduAna-Maria
RaduAna-Maria force-pushed the feat/genome-stack-1-node-scripts branch from faf76b4 to 5ae6efe Compare November 10, 2026 15:48

@danmunteanu1 danmunteanu1 left a comment

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

🟡 Commented · 1 blocker
Re-review after the rebase: one new finding on this PR; the stack-wide summary is on #4379.

@RaduAna-Maria

Copy link
Copy Markdown
Collaborator Author

Follow-up for a few earlier replies: the genome checker has since moved into uipath-planner (039379d, on #4379). The planner checks every genome it writes or changes, and re-checks after any SDD change. uipath-build-from-genome ships no script and builds the genome as written. So the builder's check verb, its context call with the designed_in* fields (N21), the blocking/report sorting of FAILs (B4), and the build-time fingerprint comparison (B8, N18) are gone. The planner's strict check now fails an unpinned or mismatched Designed in: line before a genome ever reaches a build.

@RaduAna-Maria
RaduAna-Maria force-pushed the feat/genome-stack-1-node-scripts branch from 5ae6efe to 2fc06c6 Compare November 11, 2026 09:25
RaduAna-Maria and others added 4 commits November 11, 2026 11:53
Moves the build side of uipath-genome into its own skill: execution,
parts, transactional, selector, migration and verification guides, the
genome checker as check-genome.mjs, the Studio Web override, the build
tasks and the checker tests. The registries, CODEOWNERS and activation
config switch from uipath-genome to the new skill here, once.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…ader

A blocking checker finding ended the build with a report; the planner's edit
had no row that recognised it, so the build never ran again. The build now
hands blocking findings to uipath-planner under one header the planner's edit
matches, and the report's genome edits use the same header when the user
forwards them.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Review fixes: a blocking FAIL and a missing building skill each have one
procedure (no CLI fallback); shape rule 8 and verb rule 9 honour a flow's
Decided: line; the build reads only the genome, so an edit to a designed genome
is never overridden by its SDD; Live application is asked for every genome with
UI steps, defaulting from the Deployment line; an unattended build takes the
Data Fabric default like every other question; the pack-dry-run stop becomes the
last local gate. genome-reading-guide.md keeps only what the build does with
each section, leaving the format to the planner.

Ports from main the split left out: the job's log by default (#4403), disappear
and multi-condition waits (#4435), the owning skill deciding helper workflows
(#4436), two parts at once (#4441), one package version per run and record-store
writers and readers (#4437), source fidelity and building into an existing
project (#4506), and taking the source's values (#4519).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
… reads it as a placeholder

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@RaduAna-Maria
RaduAna-Maria force-pushed the feat/genome-stack-1-node-scripts branch from 2fc06c6 to ed95c51 Compare November 11, 2026 10:00

@danmunteanu1 danmunteanu1 left a comment

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

🟡 Commented · 1 blocker. The re-review summary is on #4379.

Comment thread README.md Resolved
@RaduAna-Maria
RaduAna-Maria merged commit 075eb61 into main Nov 12, 2026
92 of 98 checks passed
@RaduAna-Maria
RaduAna-Maria deleted the feat/genome-stack-1-node-scripts branch November 12, 2026 08:55
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants