multi-agent-verify-content
Verify that another agent's (or your own) work actually landed in a squash-merge repo where PR numbers and commit SHAs are both unreliable.
Skill by kingithegreat
Chalk Labs
Need help deploying multi-agent-verify-content?
Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.
Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/kingithegreat-multi-agent-verify-content
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Use cases: writing, agent-building
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/multi-agent-verify-content
What's inside
- Skill name
multi-agent-verify-content- Description
- Verify that another agent's (or your own) work actually landed in a squash-merge repo where PR numbers and commit SHAs are both unreliable. Use after any auto-merge fires, when asked "did X ship?", when checking whether a collision with another session was resolved correctly, and before building on top of recently merged work. Covers content-grep verification, the modes.ts collision case study, and the held-runs approval loop.
- License
- MIT
- Instructions
- 62 lines in SKILL.md · view on GitHub
Files (1)
- SKILL.md
SKILL.md (first lines)
# Verify merged CONTENT, not SHAs or PR numbers
In `kingithegreat/Sadie`, auto-merge squash-merges: the original commit SHA is never an
ancestor of main, the PR number is not durable (a re-push can spawn a new PR), and several
agents ship to main within the same hour. "The PR says merged" proves nothing about what
landed.
## The verification that counts
```bash
git fetch origin
# Does the SYMBOL exist in main's version of the file?
git grep -c "'code'" origin/main -- widget/src/shared/modes.ts
git grep -c "navContext" origin/main -- widget/src/renderer/components/workspace/WorkspaceShell.tsx
```
A count ≥ 1 for every symbol your change introduced = landed. Zero = it didn't, whatever the
PR page claims. For multi-file changes, check one distinctive symbol per file.
## Collision resolution — the modes.ts case (2026-08-23)
Two sessions independently added entries to `shared/modes.ts`, `shared/navigation.ts` and
`StatusIndicator.tsx` in the same window. Both survived because each change was one line in a
shared list. What made the merge clean:
- Each addition was **additive** (append to a list/record), never a rewrite of shared lines.
- After merge, re-read the file on main and confirm BOTH entries are present before continuing
work in that area.
- If you and another session fixed the same bug independently (here: the same AppMode import
fix), expect the second one to be dropped as redundant — don't re-apply it.
## The held-runs loop (every bot PR, every time)
Bot-opened PRs park their required checks at `action_required` silently; the check never
reports and the PR reads as slow CI. After EVERY push:
```bash
gh api "repos/kingithegreat/Sadie/actions/runs?status=action_required&per_page=30" \
--jq '.workflow_runs[].id'
# then approve each id:
gh api -X POST "repos/kingithegreat/Sadie/actions/runs/<id>/approve"
```
Verify the sweep worked: `{"total_count":0,...}` means nothing is held. A non-zero total_count
with pending checks on YOUR pr = your merge is stalled on this.
## Tooling traps on this box
- `gh --jq` fails oddly with `==` inside double-quoted PowerShell strings (`function not
defined`). Prefer `gh pr checks <n>` plain text, or write JSON to a temp file and parse with
node.
- PowerShell treats git/gh stderr as errors (NativeCommandError noise) — exit code 1 often
means nothing failed; read the actual output.
- `git checkout <branch>` fails if another worktree holds that branch.
## Worktree hygiene
Ten+ worktrees accumulate fast. Before any `reset --hard` / `checkout --` / `stash`:
`git worktree list`, then per-tree `git status --porcelain`. Clean trees on merged branches are
removal candidates; dirty trees hold someone's uncommitted scratch — ask, don't delete.From kingithegreat/Sadie, MIT.
More from kingithegreat/Sadie
- 0.587score
homebot-defect-patterns
kingithegreat
Repo-specific defect patterns for HomeBot/SADIE that pass every automated check and only fail in a built, running app — dead IPC channels, relative require()…
SkillAFreshness: A★ 1Claude Code +9 - 0.587score
feature-decision-closure
kingithegreat
Close out a "should we build X" question so no future session re-litigates it — record the decision, the evidence that settled it, and the next lever if the…
SkillAFreshness: A★ 1Claude Code +9 - 0.587score
stateful-fix-patterns
kingithegreat
Failure patterns in fixes that touch persisted state, saved settings, or external CLIs — fixes that only work for new data and leave existing users broken…
SkillAFreshness: A★ 1Claude Code +9 - 0.587score
tts-engine-evaluation
kingithegreat
Settle a "should we switch TTS engines / is X better than Edge TTS" question by measuring produced audio files, not by reading library docs or trusting…
SkillAFreshness: A★ 1Claude Code +9 - 0.584score
character-art-spec
kingithegreat
What a character art kit must contain (Leila/Flappy and guests), how it is laid out, and the only approved way to generate it.
SkillAFreshness: A★ 1Claude Code +9 - 0.581score
pr-handoff-evidence
kingithegreat
Open, land and hand off a Sadie PR with acceptance evidence that another agent or Aden can check.
SkillAFreshness: A★ 1Claude Code +9 - 0.580score
homebot-dev-workflow
kingithegreat
HomeBot (kingithegreat/Sadie) build, test and landing procedures moved verbatim from AGENTS.md — the unreachable-capability defect, three traps, CI and local…
SkillAFreshness: A★ 1Claude Code +9 - 0.576score
mouth-viseme-matching
kingithegreat
Which mouth tiles exist, how audio maps to them, and on which poses a mouth may be composited.
SkillAFreshness: A★ 1Claude Code +9 - 0.576score
art-asset-naming-and-folders
kingithegreat
Where character assets live and how they must be named so the pipeline finds them.
SkillAFreshness: A★ 1Claude Code +9 - 0.576score
drive-plan-update
kingithegreat
Record verified progress in the Drive doc "HomeBot — Current Plan" without clobbering concurrent edits.
SkillAFreshness: A★ 1Claude Code +9 - 0.576score
episode-script-to-shots
kingithegreat
Turn an Ancient Pathways episode script into shots and image tickets using the existing pipeline, not a new one.
SkillAFreshness: A★ 1Claude Code +9 - 0.575score
claim-and-branch
kingithegreat
Claim a piece of HomeBot work in CLAIMS.md and start it on a safe branch/worktree so no two agents build the same thing.
SkillAFreshness: A★ 1Claude Code +9
Signals
- Stars
- 1
- Forks
- 0
- Open issues
- 1
- Last commit (this folder)
- Oct 20, 2026 (11 days ago)
- Last release
- Dec 14, 2025 (10 months ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- MIT
- Repository
- github.com/kingithegreat/Sadie
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from kingithegreat
- 0.587score
stateful-fix-patterns
kingithegreat
Failure patterns in fixes that touch persisted state, saved settings, or external CLIs — fixes that only work for new data and leave existing users broken…
SkillAFreshness: A★ 1Claude Code +9 - 0.587score
feature-decision-closure
kingithegreat
Close out a "should we build X" question so no future session re-litigates it — record the decision, the evidence that settled it, and the next lever if the…
SkillAFreshness: A★ 1Claude Code +9 - 0.587score
homebot-defect-patterns
kingithegreat
Repo-specific defect patterns for HomeBot/SADIE that pass every automated check and only fail in a built, running app — dead IPC channels, relative require()…
SkillAFreshness: A★ 1Claude Code +9 - 0.587score
tts-engine-evaluation
kingithegreat
Settle a "should we switch TTS engines / is X better than Edge TTS" question by measuring produced audio files, not by reading library docs or trusting…
SkillAFreshness: A★ 1Claude Code +9 - 0.584score
character-art-spec
kingithegreat
What a character art kit must contain (Leila/Flappy and guests), how it is laid out, and the only approved way to generate it.
SkillAFreshness: A★ 1Claude Code +9 - 0.581score
pr-handoff-evidence
kingithegreat
Open, land and hand off a Sadie PR with acceptance evidence that another agent or Aden can check.
SkillAFreshness: A★ 1Claude Code +9
Done for you
Rather not set it up yourself?
Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.
See how we build AI agents →Read the case studies
Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.
- launches shipped
- 50+
- launches shipped
- media placements
- 500+
- media placements
- founders on every account
- 2
- founders on every account
Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block
Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.
Get multi-agent-verify-content running in your business
Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.
Cite this page
multi-agent-verify-content skill. Agent Mart. https://agentmart.dev/skills/kingithegreat-multi-agent-verify-content (last updated 2026-10-24)