Skip to content
Open
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
28 changes: 14 additions & 14 deletions action.yml
Original file line number Diff line number Diff line change
Expand Up @@ -1445,7 +1445,7 @@ runs:
- name: Run Claude Code (attempt 1)
id: attempt_1
continue-on-error: true
uses: anthropics/claude-code-action@e8c2d7c16c018cf1e694711c1c07a5f5db2b5eb1 # v1.0.208
uses: anthropics/claude-code-action@d75b94d5ad426cb8546e6628b6f5f19b84e5cce1 # v1.0.216
with:
prompt: ${{ steps.prompt.outputs.text }}
claude_args: ${{ inputs.claude_args || steps.resume.outputs.claude_args || steps.args.outputs.claude_args }}
Expand Down Expand Up @@ -1523,7 +1523,7 @@ runs:
id: attempt_1_retry
if: steps.retry_gate_1.outputs.should_retry == 'true'
continue-on-error: true
uses: anthropics/claude-code-action@e8c2d7c16c018cf1e694711c1c07a5f5db2b5eb1 # v1.0.208
uses: anthropics/claude-code-action@d75b94d5ad426cb8546e6628b6f5f19b84e5cce1 # v1.0.216
with:
prompt: ${{ steps.prompt.outputs.text }}
# No steps.resume.outputs.claude_args here -- see attempt_2's own comment further up: a same-credential retry within this same job run must always start fresh, never resume a session the failed first try in this run may have already written to.
Expand Down Expand Up @@ -1570,7 +1570,7 @@ runs:
id: attempt_2
if: steps.credentials.outputs.cred_2_token != '' && !(steps.attempt_1.outcome == 'success' || steps.attempt_1_retry.outcome == 'success')
continue-on-error: true
uses: anthropics/claude-code-action@e8c2d7c16c018cf1e694711c1c07a5f5db2b5eb1 # v1.0.208
uses: anthropics/claude-code-action@d75b94d5ad426cb8546e6628b6f5f19b84e5cce1 # v1.0.216
with:
prompt: ${{ steps.prompt.outputs.text }}
claude_args: ${{ inputs.claude_args || steps.args.outputs.claude_args }}
Expand Down Expand Up @@ -1648,7 +1648,7 @@ runs:
id: attempt_2_retry
if: steps.retry_gate_2.outputs.should_retry == 'true'
continue-on-error: true
uses: anthropics/claude-code-action@e8c2d7c16c018cf1e694711c1c07a5f5db2b5eb1 # v1.0.208
uses: anthropics/claude-code-action@d75b94d5ad426cb8546e6628b6f5f19b84e5cce1 # v1.0.216
with:
prompt: ${{ steps.prompt.outputs.text }}
# No steps.resume.outputs.claude_args here -- see attempt_2's own comment further up: a same-credential retry within this same job run must always start fresh, never resume a session the failed first try in this run may have already written to.
Expand Down Expand Up @@ -1695,7 +1695,7 @@ runs:
id: attempt_3
if: steps.credentials.outputs.cred_3_token != '' && !(steps.attempt_2.outcome == 'success' || steps.attempt_2_retry.outcome == 'success')
continue-on-error: true
uses: anthropics/claude-code-action@e8c2d7c16c018cf1e694711c1c07a5f5db2b5eb1 # v1.0.208
uses: anthropics/claude-code-action@d75b94d5ad426cb8546e6628b6f5f19b84e5cce1 # v1.0.216
with:
prompt: ${{ steps.prompt.outputs.text }}
claude_args: ${{ inputs.claude_args || steps.args.outputs.claude_args }}
Expand Down Expand Up @@ -1773,7 +1773,7 @@ runs:
id: attempt_3_retry
if: steps.retry_gate_3.outputs.should_retry == 'true'
continue-on-error: true
uses: anthropics/claude-code-action@e8c2d7c16c018cf1e694711c1c07a5f5db2b5eb1 # v1.0.208
uses: anthropics/claude-code-action@d75b94d5ad426cb8546e6628b6f5f19b84e5cce1 # v1.0.216
with:
prompt: ${{ steps.prompt.outputs.text }}
# No steps.resume.outputs.claude_args here -- see attempt_2's own comment further up: a same-credential retry within this same job run must always start fresh, never resume a session the failed first try in this run may have already written to.
Expand Down Expand Up @@ -1820,7 +1820,7 @@ runs:
id: attempt_4
if: steps.credentials.outputs.cred_4_token != '' && !(steps.attempt_3.outcome == 'success' || steps.attempt_3_retry.outcome == 'success')
continue-on-error: true
uses: anthropics/claude-code-action@e8c2d7c16c018cf1e694711c1c07a5f5db2b5eb1 # v1.0.208
uses: anthropics/claude-code-action@d75b94d5ad426cb8546e6628b6f5f19b84e5cce1 # v1.0.216
with:
prompt: ${{ steps.prompt.outputs.text }}
claude_args: ${{ inputs.claude_args || steps.args.outputs.claude_args }}
Expand Down Expand Up @@ -1898,7 +1898,7 @@ runs:
id: attempt_4_retry
if: steps.retry_gate_4.outputs.should_retry == 'true'
continue-on-error: true
uses: anthropics/claude-code-action@e8c2d7c16c018cf1e694711c1c07a5f5db2b5eb1 # v1.0.208
uses: anthropics/claude-code-action@d75b94d5ad426cb8546e6628b6f5f19b84e5cce1 # v1.0.216
with:
prompt: ${{ steps.prompt.outputs.text }}
# No steps.resume.outputs.claude_args here -- see attempt_2's own comment further up: a same-credential retry within this same job run must always start fresh, never resume a session the failed first try in this run may have already written to.
Expand Down Expand Up @@ -1945,7 +1945,7 @@ runs:
id: attempt_5
if: steps.credentials.outputs.cred_5_token != '' && !(steps.attempt_4.outcome == 'success' || steps.attempt_4_retry.outcome == 'success')
continue-on-error: true
uses: anthropics/claude-code-action@e8c2d7c16c018cf1e694711c1c07a5f5db2b5eb1 # v1.0.208
uses: anthropics/claude-code-action@d75b94d5ad426cb8546e6628b6f5f19b84e5cce1 # v1.0.216
with:
prompt: ${{ steps.prompt.outputs.text }}
claude_args: ${{ inputs.claude_args || steps.args.outputs.claude_args }}
Expand Down Expand Up @@ -2023,7 +2023,7 @@ runs:
id: attempt_5_retry
if: steps.retry_gate_5.outputs.should_retry == 'true'
continue-on-error: true
uses: anthropics/claude-code-action@e8c2d7c16c018cf1e694711c1c07a5f5db2b5eb1 # v1.0.208
uses: anthropics/claude-code-action@d75b94d5ad426cb8546e6628b6f5f19b84e5cce1 # v1.0.216
with:
prompt: ${{ steps.prompt.outputs.text }}
# No steps.resume.outputs.claude_args here -- see attempt_2's own comment further up: a same-credential retry within this same job run must always start fresh, never resume a session the failed first try in this run may have already written to.
Expand Down Expand Up @@ -2335,7 +2335,7 @@ runs:
- name: Wrap up on turn limit
id: wrap-up
if: ${{ steps.detect-turn-limit.outputs.hit_limit == 'true' }}
uses: anthropics/claude-code-action@e8c2d7c16c018cf1e694711c1c07a5f5db2b5eb1 # v1.0.208
uses: anthropics/claude-code-action@d75b94d5ad426cb8546e6628b6f5f19b84e5cce1 # v1.0.216
with:
prompt: |
You may have been cut off by a turn limit in your previous run on this pull request or issue. Reply now, briefly and honestly: if your work is complete, say "Done." If you were interrupted, summarise what you found, what you completed, and what remains to be done. Do not start new work — this is a closing reply, not a continuation.
Expand Down Expand Up @@ -2383,7 +2383,7 @@ runs:
id: escalate
if: ${{ inputs.mode == 'review' && inputs.escalate_findings_to_issues == 'true' && steps.wrap-up.outcome == 'failure' }}
continue-on-error: true
uses: anthropics/claude-code-action@e8c2d7c16c018cf1e694711c1c07a5f5db2b5eb1 # v1.0.208
uses: anthropics/claude-code-action@d75b94d5ad426cb8546e6628b6f5f19b84e5cce1 # v1.0.216
with:
prompt: |
Your last reply in this session could not be posted where it was meant to go -- most likely because that write was blocked, or because the pull request had already merged or closed underneath you. This is a further, final pass: decide whether anything you already found is worth a standing record even though your normal output channel is unavailable.
Expand Down Expand Up @@ -2493,7 +2493,7 @@ runs:
id: apply-fixes
if: ${{ steps.fix-args.outputs.can_push == 'true' }}
continue-on-error: true
uses: anthropics/claude-code-action@e8c2d7c16c018cf1e694711c1c07a5f5db2b5eb1 # v1.0.208
uses: anthropics/claude-code-action@d75b94d5ad426cb8546e6628b6f5f19b84e5cce1 # v1.0.216
with:
prompt: |
You have just submitted your review of pull request #${{ inputs.pr_number || github.event.pull_request.number }} in ${{ github.repository }}, in this same session. This is a separate, final pass with a different job: apply the fixes you are confident about, and commit them. It does not change the review you already submitted, and you must not submit another one.
Expand Down Expand Up @@ -2613,7 +2613,7 @@ runs:
id: structured-summary
if: ${{ inputs.mode == 'review' && inputs.structured_review_summary == 'true' && steps.claude.outputs.session_id != '' && steps.detect-turn-limit.outputs.hit_limit != 'true' }}
continue-on-error: true
uses: anthropics/claude-code-action@e8c2d7c16c018cf1e694711c1c07a5f5db2b5eb1 # v1.0.208
uses: anthropics/claude-code-action@d75b94d5ad426cb8546e6628b6f5f19b84e5cce1 # v1.0.216
with:
prompt: |
You just finished reviewing pull request #${{ inputs.pr_number || github.event.pull_request.number }} in ${{ github.repository }} in this same session. Do not review anything further, and do not re-judge your findings from memory.
Expand Down
Loading