fix(grant-scout): harden Linear access, dedup, and anti-fabrication (supersedes #19)#25
Open
Oba-One wants to merge 1 commit into
Open
fix(grant-scout): harden Linear access, dedup, and anti-fabrication (supersedes #19)#25Oba-One wants to merge 1 commit into
Oba-One wants to merge 1 commit into
Conversation
Addresses the 2026-05-28 run failures (Linear "unreachable" + the fabricated "EF ESP passkey-support RFP"). Eight coordinated fixes: - Linear via the GraphQL API + LINEAR_API_KEY as the canonical headless-safe path (an OAuth connector that lapses can't re-consent in a cron); fail-closed when Linear is unreachable instead of scouting with no dedup. - Read funding:* workspace-wide; create non-awarded grants in Research / Grant Scouting (the pipeline moved off the Product team on 2026-05-30); awards graduate to Product. Dedup keyed on funder URL, not just title. - Phase 2 existence gate: every opportunity must resolve to a funder page fetched this run; unconfirmable -> "Unverified leads", never a prospect. - Reframe the 6-8 "NEW" quota to breadth-of-scanning; a quiet week beats a fabricated one. - Discord output: bare grant URLs (previews, top <=5 by fit), Linear links grouped and <>-suppressed (private workspace, no previews). - Run-start preflight + a health line in the memo. - Folds in #19: removes the "Evidence Commons" false-prior-grant claim and the hardcoded "13 live gardens"; adds a verify-claims-vs-primary-source guardrail. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
📝 WalkthroughSummary by CodeRabbit
WalkthroughThe PR updates the Guild Grant Scout routine specification to add fail-closed Linear integration, enforce existence verification before scoring or surfacing opportunities, restructure Linear issue routing and workspace-wide deduping, and standardize Discord/memo output with expanded safety guardrails. ChangesGuild Grant Scout Routine Enhancement
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~25 minutes Poem
🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 2
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@routines/claude/guild-grant-scout.md`:
- Line 54: The line describing "Green Goods" contains the guild-wide phrase
"Season One" which violates the repo documentation policy; edit the markdown in
guild-grant-scout.md to remove the "Season One" framing (e.g., change "Season
One pilot live on Arbitrum — pull current garden / gardener / operator counts at
runtime per the Phase 2/3 metrics step; cite 'as of {date}'" to a neutral phrase
like "pilot live on Arbitrum — pull current garden / gardener / operator counts
at runtime per the Phase 2/3 metrics step; cite 'as of {date}'" or otherwise
omit any "Season" wording) so the entry for "**Green Goods**" no longer contains
guild-wide "Season" language.
- Around line 43-44: Update the cross-file lifecycle contract so all referenced
docs and routine catalog entries reflect the Research-first routing (non-awarded
issues in Research's "Grant Scouting" project with canonical funding:* labels,
awarded grants moved to Product), or explicitly mark this spec as superseding
older Product-centered docs; search for references to "funding:*", "Grant
Scouting", "LINEAR_API_KEY", and any doc/routine entries that describe grant
lifecycle or Linear routing and change their text/headers to match this new
Research-first behavior (or add a clear supersession note), and ensure the
fail-closed Phase 0 / Linear-as-canonical guidance is consistently stated across
the routine catalog and related docs.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
Run ID: cf3c8787-0908-4d31-bd22-99f8c9126e46
📒 Files selected for processing (1)
routines/claude/guild-grant-scout.md
📜 Review details
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (1)
- GitHub Check: Analyze (javascript-typescript)
🧰 Additional context used
📓 Path-based instructions (1)
**/*.{md,yml}
📄 CodeRabbit inference engine (AGENTS.md)
**/*.{md,yml}: Do not reintroduce the former dev-guild PM URL or old migration language for project-management state
Do not introduce guild-wide 'Season' framing in documentation
Do not describe the dev guild as the entire Greenpill Network in documentation
Avoid growth-hacking and trend-chasing voice in documentation: no FOMO, virality, leaderboards, streaks, or 'AI-first' framing
Files:
routines/claude/guild-grant-scout.md
🧠 Learnings (6)
📚 Learning: 2026-05-31T05:16:36.572Z
Learnt from: CR
Repo: greenpill-dev-guild/.github PR: 0
File: CLAUDE.md:0-0
Timestamp: 2026-05-31T05:16:36.572Z
Learning: Treat `package.json`, lockfiles, package-manager config, `.github/workflows/**`, `AGENTS.md`, `CLAUDE.md`, `.codex/**`, and `.claude/**` as security-sensitive surfaces and call out any changes in final summaries
Applied to files:
routines/claude/guild-grant-scout.md
📚 Learning: 2026-05-31T05:16:51.577Z
Learnt from: CR
Repo: greenpill-dev-guild/.github PR: 0
File: AGENTS.md:0-0
Timestamp: 2026-05-31T05:16:51.577Z
Learning: Applies to {package.json,*lock*,*.lock.json,.npmrc,.yarnrc*,.github/workflows/**,AGENTS.md,CLAUDE.md,.codex/**,.claude/**} : Treat `package.json`, lockfiles, package-manager config, `.github/workflows/**`, `AGENTS.md`, `CLAUDE.md`, `.codex/**`, and `.claude/**` as security-sensitive surfaces and call out any changes in final summaries
Applied to files:
routines/claude/guild-grant-scout.md
📚 Learning: 2026-05-31T05:16:36.572Z
Learnt from: CR
Repo: greenpill-dev-guild/.github PR: 0
File: CLAUDE.md:0-0
Timestamp: 2026-05-31T05:16:36.572Z
Learning: Name Linear as the project-management source of truth; use GitHub for public execution/RFCs/code review, Drive for memos/evidence, and Discord/Telegram/calls for discussion
Applied to files:
routines/claude/guild-grant-scout.md
📚 Learning: 2026-05-31T05:16:51.577Z
Learnt from: CR
Repo: greenpill-dev-guild/.github PR: 0
File: AGENTS.md:0-0
Timestamp: 2026-05-31T05:16:51.577Z
Learning: Linear is the project-management source of truth; GitHub owns public execution/RFCs/code review; Drive owns memos and evidence; Discord/Telegram/calls own discussion
Applied to files:
routines/claude/guild-grant-scout.md
📚 Learning: 2026-05-31T05:16:36.572Z
Learnt from: CR
Repo: greenpill-dev-guild/.github PR: 0
File: CLAUDE.md:0-0
Timestamp: 2026-05-31T05:16:36.572Z
Learning: Keep the guild voice grounded in public goods, transparency, and building in public; avoid hype framing
Applied to files:
routines/claude/guild-grant-scout.md
📚 Learning: 2026-05-31T05:16:51.577Z
Learnt from: CR
Repo: greenpill-dev-guild/.github PR: 0
File: AGENTS.md:0-0
Timestamp: 2026-05-31T05:16:51.577Z
Learning: Applies to **/{CONTRIBUTING.md,GOVERNANCE.md,SECURITY.md,SUPPORT.md} : Treat contact emails, disclosure SLAs, payout timing, and governance claims as sensitive public commitments and do not change them casually
Applied to files:
routines/claude/guild-grant-scout.md
🪛 LanguageTool
routines/claude/guild-grant-scout.md
[style] ~78-~78: Consider an alternative for the overused word “exactly”.
Context: ...ntly misses half the pipeline — this is exactly what re-surfaces already-tracked progra...
(EXACTLY_PRECISELY)
[uncategorized] ~473-~473: The official name of this software platform is spelled with a capital “H”.
Context: ...nt lifecycle Issues anywhere else — not .github (no GitHub Issues, ever), not project ...
(GITHUB)
Comment on lines
+43
to
+44
| - **Linear is the canonical surface for grant lifecycle.** Non-awarded grants (prospect / drafting / submitted) live as Issues in the **Research team's `Grant Scouting` project**, tracked with the canonical `funding:*` lifecycle labels. **Awarded grants graduate** to the **Product team** into a bounded award/delivery project when delivery, reporting, compliance, or funder follow-through needs project-level management. Because awards live in Product and everything else in Research, **always read `funding:*` workspace-wide — never scope a read to a single team** (a team-scoped read silently misses half the pipeline and causes already-tracked programs to re-surface as "NEW"). Resolve team/project/label IDs by name at the start of every run. | ||
| - **Reach Linear via the GraphQL API with `LINEAR_API_KEY` — this is the canonical, headless-safe path.** A scheduled run must not depend on an interactive OAuth connector that can silently lapse between runs. Endpoint `https://api.linear.app/graphql`; headers `Authorization: $LINEAR_API_KEY` and `Content-Type: application/json`; body `{"query": "..."}` (reads) or `{"query": "mutation ..."}` (writes), e.g. via `curl`. The Linear MCP connector may be used when present, but never rely on it. If `LINEAR_API_KEY` is unset or the API is unreachable, you have lost the canonical dedup + write surface — follow the **fail-closed rule in Phase 0** (do not scout; post one status line; exit). Never scout without Linear. |
There was a problem hiding this comment.
Sync cross-file Linear lifecycle contracts with this new Research-first routing.
These lines change the lifecycle contract (non-awarded in Research, awards in Product), but referenced docs still describe lifecycle issues as Product-centered. Please update the related docs/routine catalog in the same PR (or explicitly mark this spec as superseding them) to avoid operator drift and conflicting automation assumptions.
Based on learnings, “Linear is the project-management source of truth; GitHub owns public execution/RFCs/code review; Drive owns memos and evidence; Discord/Telegram/calls own discussion”.
Also applies to: 271-272, 275-277, 281-281
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@routines/claude/guild-grant-scout.md` around lines 43 - 44, Update the
cross-file lifecycle contract so all referenced docs and routine catalog entries
reflect the Research-first routing (non-awarded issues in Research's "Grant
Scouting" project with canonical funding:* labels, awarded grants moved to
Product), or explicitly mark this spec as superseding older Product-centered
docs; search for references to "funding:*", "Grant Scouting", "LINEAR_API_KEY",
and any doc/routine entries that describe grant lifecycle or Linear routing and
change their text/headers to match this new Research-first behavior (or add a
clear supersession note), and ensure the fail-closed Phase 0 /
Linear-as-canonical guidance is consistently stated across the routine catalog
and related docs.
| What this guild is uniquely positioned to win funding for: | ||
|
|
||
| - **Green Goods** — offline-first regenerative documentation platform. Passkey auth, EAS attestations, Arbitrum production deployment with real users (Season One: 13 live gardens), Envio indexer. Strongest grant evidence in the guild — production code + production users. | ||
| - **Green Goods** — offline-first regenerative documentation platform. Passkey auth, EAS attestations, Arbitrum production deployment with real users (Season One pilot live on Arbitrum — pull current garden / gardener / operator counts at runtime per the Phase 2/3 metrics step; cite "as of {date}", never a hardcoded number), Envio indexer. Strongest grant evidence in the guild — production code + production users. |
There was a problem hiding this comment.
Remove guild-wide “Season One” framing from this line.
This introduces “Season One” wording in documentation, which conflicts with the repo documentation policy for .md files.
As per coding guidelines, “Do not introduce guild-wide 'Season' framing in documentation”.
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@routines/claude/guild-grant-scout.md` at line 54, The line describing "Green
Goods" contains the guild-wide phrase "Season One" which violates the repo
documentation policy; edit the markdown in guild-grant-scout.md to remove the
"Season One" framing (e.g., change "Season One pilot live on Arbitrum — pull
current garden / gardener / operator counts at runtime per the Phase 2/3 metrics
step; cite 'as of {date}'" to a neutral phrase like "pilot live on Arbitrum —
pull current garden / gardener / operator counts at runtime per the Phase 2/3
metrics step; cite 'as of {date}'" or otherwise omit any "Season" wording) so
the entry for "**Green Goods**" no longer contains guild-wide "Season" language.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Why
The 2026-05-28 run failed two ways: "Linear unreachable" (again) and a fabricated opportunity — "EF ESP passkey-support RFP", which is a real-but-archived program (RESR-37) re-surfaced as NEW with an invented RFP framing. Root-caused to this spec. Five root causes → eight coordinated fixes (one file:
routines/claude/guild-grant-scout.md).What changed
Linear connectivity
LINEAR_API_KEY(headless-safe). The OAuth MCP connector lapses between runs and a cron can't re-consent, so it's now optional, not the dependency.Correct Linear routing (independent of connectivity)
funding:*label (was: Product team only). The pipeline moved to Research /Grant Scoutingon 2026-05-30 and awards live in Product, so the old single-team query silently missed half the pipeline.Anti-fabrication
⚠️ unverified — human confirm(so bot-blocked climate / Africa / gov portals aren't punished); no URL → "Unverified leads". Rolling / always-open counts as open.Output
<>-suppressed (private workspace, no previews). Adds a 🔍 Unverified-leads section and a run-start health line in the memo.LINEAR_API_KEYin theguild-routinesenv — a Linear personal API key (header isAuthorization: <key>, noBearer). Until it's set, every run correctly fail-closes to a no-op status post.The thin-wrapper trigger re-reads this file each run, so it ships on merge with no trigger reconfig.
🤖 Generated with Claude Code