fix(#1030): use TRIGGER_ENTITY_URL for Jira forge ISSUE_URL mapping

[fullsend-ai-coder]

Summary

• Replace JIRA_ISSUE_URL with TRIGGER_ENTITY_URL in the forge.jira section of harness/triage.yaml (env.runner and env.sandbox blocks) and in env/jira/triage.env.
• JIRA_ISSUE_URL was never set by the dispatch workflow, so the Jira forge ISSUE_URL never resolved. TRIGGER_ENTITY_URL is the new forge-agnostic variable added in chore(dispatch): provide FULLSEND_WORK_ITEM_URL in addition to GITHUB_ISSUE_URL fullsend#6611.

Testing

• Verified YAML validity of harness/triage.yaml after the change.
• Ran lint-agent-docs hook — all harness doc references and structure checks pass.
• Ran make test — all test suites pass (16 pre-existing failures in post-retro-test.sh are unrelated and reproduce identically on main).

---

Closes #1030

Post-script verification

• Branch is not main/master (agent/1030-jira-trigger-entity-url)
• Secret scan passed (gitleaks — b9c0745580084af09d1f1e2df8448d1d478458c6..HEAD)
• PR body secret scan passed (gitleaks — no-git)

[fullsend-ai-review]

🤖 Finished Review · ✅ Success · Started 8:44 PM UTC · Completed 8:59 PM UTC

Commit: 083d9b2 · View workflow run →

Runtime: claude · Model: opus → claude-opus-4-6 · Effort: high · Cost: $2.60

[fullsend-ai-review]

Risk Assessment: moderate (2/5)

Details

Small, well-scoped env var rename (JIRA_ISSUE_URL to TRIGGER_ENTITY_URL) across 4 config/doc files by a trusted bot author, with clear issue alignment and easy rollback; elevated git churn on some files is typical for frequently-updated harness configs.

Previous run

Risk Assessment: moderate (2/5)

Details

Small config-only change from a bot replacing a non-functional env var (JIRA_ISSUE_URL) with the correct one (TRIGGER_ENTITY_URL); risk is moderate due to harness/triage.yaml being a high-churn file with frequent fix/revert history, but the change itself is minimal and well-scoped to the linked issue.

[fullsend-ai-review]

Review

Findings

Medium

• [protected-path] harness/triage.yaml — This PR modifies harness/triage.yaml, which is under the protected harness/ path. The change is authorized by issue Update the forge.jira env var to use TRIGGER_ENTITY_URL instead of JIRA_ISSUE_URL when setting ISSUE_URL #1030 and the rationale is clearly explained. Human approval is always required for protected-path changes.

Low

• [technical-inaccuracy] docs/triage.md:228 — After this PR, line 228 reads: "GITHUB_ISSUE_URL (and its GITLAB_ISSUE_URL / TRIGGER_ENTITY_URL equivalents) remain the workflow-level inputs per forge". The word "equivalents" is misleading: TRIGGER_ENTITY_URL is the forge-agnostic replacement for JIRA_ISSUE_URL, not another forge-specific input like GITLAB_ISSUE_URL. Grouping all three as peer "equivalents" obscures the architectural distinction.
  Remediation: Rewrite the parenthetical to reflect the new hierarchy, e.g.: "GITHUB_ISSUE_URL (and its GITLAB_ISSUE_URL equivalent for GitLab) remain the workflow-level inputs per forge; Jira now uses the forge-agnostic TRIGGER_ENTITY_URL instead of the former JIRA_ISSUE_URL."

Previous run

Review

Findings

Medium

• [stale-reference] LOCAL.md:133 — LOCAL.md still instructs users to export JIRA_ISSUE_URL=... for local Jira triage setup. After this PR, the harness reads TRIGGER_ENTITY_URL instead, so users following this guide will get an empty ISSUE_URL.
  Remediation: Update LOCAL.md (lines ~78 and ~133) to reference TRIGGER_ENTITY_URL instead of JIRA_ISSUE_URL.

• [protected-path] harness/triage.yaml — This PR modifies harness/triage.yaml, which is under the protected harness/ path. The change is authorized by issue Update the forge.jira env var to use TRIGGER_ENTITY_URL instead of JIRA_ISSUE_URL when setting ISSUE_URL #1030 and the rationale is clearly explained. Human approval is always required for protected-path changes.

Low

• [stale-reference] docs/triage.md:193 — docs/triage.md documents JIRA_ISSUE_URL as a required Jira setup variable. After this PR, the harness reads TRIGGER_ENTITY_URL instead for the Jira forge's ISSUE_URL mapping.
  Remediation: Update the Jira variable table entry to reference TRIGGER_ENTITY_URL.

[ralphbean]

/fs-fix fix the LOCAL.md issue

[ralphbean]

/fs-fix actually, fix both LOCAL.md and docs/triage.md - just replace the var name.

[fullsend-ai-coder]

🤖 Finished Fix · ✅ Success · Started 9:37 PM UTC · Completed 9:42 PM UTC

Commit: 083d9b2 · View workflow run →

Runtime: claude · Model: opus → claude-opus-4-6 · Effort: high · Cost: $1.23

[fullsend-ai-coder]

🔧 Fix agent — iteration 1 (human-triggered)

Replaced all 4 occurrences of JIRA_ISSUE_URL with TRIGGER_ENTITY_URL in LOCAL.md and docs/triage.md, as requested by human instruction. This completes the variable rename across documentation to match the config changes already in the PR.

Fixed (4):

1. Replace JIRA_ISSUE_URL in docs/triage.md Jira setup table (docs/triage.md): Replaced JIRA_ISSUE_URL with TRIGGER_ENTITY_URL in the Jira env vars table (line 193)
2. Replace JIRA_ISSUE_URL in docs/triage.md migration notes (docs/triage.md): Replaced JIRA_ISSUE_URL with TRIGGER_ENTITY_URL in the migration notes section (line 228)
3. Replace JIRA_ISSUE_URL in LOCAL.md harness mapping explanation (LOCAL.md): Replaced JIRA_ISSUE_URL with TRIGGER_ENTITY_URL in the var mapping explanation (line 78)
4. Replace JIRA_ISSUE_URL in LOCAL.md Jira example export (LOCAL.md): Replaced JIRA_ISSUE_URL with TRIGGER_ENTITY_URL in the example export command (line 133)

Tests: passed

Next steps:

• /fs-review — request a re-review of the changes
• /fs-fix <your instruction> — run another fix pass with specific guidance
• Push commits directly — review re-runs automatically on push
  _{Updated by fullsend fix agent}

[fullsend-ai-review]

🤖 Finished Review · ✅ Success · Started 9:44 PM UTC · Completed 9:58 PM UTC

Commit: 7a84d59 · View workflow run →

Runtime: claude · Model: opus → claude-opus-4-6 · Effort: high · Cost: $3.05

[fullsend-ai-review]

See the review comment for full details.

[fullsend-ai-retro]

🤖 Finished Retro · ✅ Success · Started 11:33 PM UTC · Completed 11:42 PM UTC

Commit: 7a84d59 · View workflow run →

Runtime: claude · Model: opus → claude-opus-4-6 · Effort: high · Cost: $3.86

[fullsend-ai-retro]

Retro: PR #1031 — fix(#1030): use TRIGGER_ENTITY_URL for Jira forge ISSUE_URL mapping

Workflow quality: Good. The pipeline completed in ~3 hours at $6.88 total cost. The review agent caught genuine gaps in the code agent's work, a human-triggered fix cycle addressed them, and the re-review surfaced a remaining minor semantic issue that the human reviewer accepted.

Timeline

1. Issue Update the forge.jira env var to use TRIGGER_ENTITY_URL instead of JIRA_ISSUE_URL when setting ISSUE_URL #1030 opened by ralphbean (20:32Z) — clear, prescriptive task: replace JIRA_ISSUE_URL with TRIGGER_ENTITY_URL in the Jira forge triage config.
2. Triage agent failed (run 32895889425, 20:32Z) — did not block downstream agents.
3. Code agent committed 083d9b2 (20:40Z) — changed harness/triage.yaml and env/jira/triage.env but missed stale JIRA_ISSUE_URL references in LOCAL.md and docs/triage.md.
4. Review run 1 (32896800620, $2.60) — correctly flagged the two stale-reference findings plus a protected-path flag for harness/triage.yaml.
5. Human /fs-fix (21:35Z) — ralphbean directed the fix agent to update both doc files.
6. Fix agent (32901833119, $1.23) — pushed 7a84d59 with the doc updates.
7. Review run 2 (32902468426, $3.05) — confirmed stale-reference findings resolved; surfaced a low-severity technical-inaccuracy where the mechanical replacement grouped the forge-agnostic TRIGGER_ENTITY_URL as an equivalent of forge-specific variables.
8. Merged (23:31Z) — ralphbean approved and merged with the low-severity finding unresolved.

Rework analysis

The single fix iteration was caused by the code agent not performing a repo-wide search for JIRA_ISSUE_URL before committing. A grep would have found all 4 affected files. This gap added one fix cycle ($1.23) and a re-review ($3.05) — roughly 60% of the total pipeline cost.

Existing issue coverage

All improvement opportunities identified in this retro are already tracked by open issues:

• Code agent should do repo-wide grep for old identifiers during rename/refactor tasks #656 (code agent repo-wide grep for old identifiers during rename/refactor) — directly addresses the root cause of the fix cycle. This PR is additional evidence: the code agent replaced JIRA_ISSUE_URL in config files but did not grep for the old identifier in docs.
• Code agent should proactively update documentation when changing user-facing behavior #944 (code agent proactive documentation updates) — overlaps with Code agent should do repo-wide grep for old identifiers during rename/refactor tasks #656 for this case; the code agent changed a user-facing env var without updating the local-setup guide and docs.
• Fix agent should check for documentation consistency after structural changes #906 (fix agent documentation consistency after structural changes) — the fix agent's mechanical find-and-replace at docs/triage.md:228 introduced a semantic inaccuracy by grouping a forge-agnostic variable with forge-specific peers.
• Fix agent should enumerate and address all review findings, not silently drop some #864 / Fix agent should verify its changes resolve the triggering review findings before committing #665 (fix agent enumerating and verifying all review findings) — the fix agent addressed the stale-reference findings but the resulting text had a new (minor) issue the re-review caught.

Review quality

The review agent performed well across both runs. It identified actionable stale-reference gaps in run 1 and a legitimate (if low-severity) semantic issue in run 2. The human reviewer's approval was appropriate — the review agent had already provided thorough analysis, and the remaining finding was low-severity with no functional impact.

No new proposals are warranted — existing issues cover the identified improvement opportunities.