Skip to content

fix: align H1 headings with frontmatter title convention #36

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
waynesun09 merged 6 commits into from
Jul 2, 2026
Merged

fix: align H1 headings with frontmatter title convention #36

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
69e8ece
fix: align H1 headings with frontmatter title convention
waynesun09 Jul 2, 2026
d9e7732
fix(0023): apply experiment conventions to review-cache-experiment
waynesun09 Jul 2, 2026
7cc6000
feat: add H1-matches-frontmatter-title lint check
waynesun09 Jul 2, 2026
59cfec8
docs(0006): rename EXPERIMENT.md to README.md
waynesun09 Jul 2, 2026
de0b2c6
style(0023): apply pre-commit auto-fixes
waynesun09 Jul 2, 2026
05bd450
fix: address review findings from PR #36
waynesun09 Jul 2, 2026
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
2 changes: 1 addition & 1 deletion 0001-agent-outage-fire-drill.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,7 +6,7 @@ topics:
- incident-response
---

# Experiment 003: Agent Outage Fire Drill
# 1. Agent outage fire drill

**Date:** 2026-03-16
**Status:** Proposed
Expand Down
2 changes: 1 addition & 1 deletion 0002-claude-scanner/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,7 +6,7 @@ topics:
- tooling
---

# Experiment 002: Claude-based ADR Drift Scanner
# 2. Claude-based ADR drift scanner

@qodo-code-review qodo-code-review Bot Jul 2, 2026

Copy link
Copy Markdown

Choose a reason for hiding this comment

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

Informational

2. Stale experiment link labels 🐞 Bug ⚙ Maintainability

Several docs still refer to experiments using the old label format (e.g., “Experiment 002”) even
though the experiment pages are now titled as “2. …”. This leaves cross-references that no longer
match the destination page titles/H1s.
Agent Prompt 
## Issue description
After renaming experiment entry H1s/titles to the `N. Title` convention, multiple in-repo references still use the legacy `Experiment 00N` wording. This makes cross-references inconsistent with the new canonical experiment titles.

## Issue Context
Examples include relationship sections and comparison headings that still say “Experiment 002”, “Experiment 003”, etc., while the target pages now start with “2.”, “3.”, etc.

## Fix Focus Areas
- 0001-agent-outage-fire-drill.md[100-103]
- 0003-scanner/README.md[23-29]
- 0004-meta-loop-self-improving-engine/README.md[140-155]
- 0011-integration-service-design-drift/README.md[257-264]

ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools


A shell script that uses `claude` CLI to analyze Tekton tasks for ADR compliance. No hardcoded rules, no config files — all intelligence comes from claude's comprehension of the ADR text.

Expand Down
2 changes: 1 addition & 1 deletion 0003-scanner/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,7 +6,7 @@ topics:
- tooling
---

# Experiment 001: ADR-0046 Drift Scanner
# 3. ADR-0046 drift scanner

A Python CLI tool that detects Tekton task steps drifting from [ADR-0046](https://github.com/konflux-ci/architecture/blob/main/ADR/0046-common-task-runner-image.md) (common task runner image) by comparing step images against a config-driven allowlist.

Expand Down
2 changes: 1 addition & 1 deletion 0004-meta-loop-self-improving-engine/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7,7 +7,7 @@ topics:
- evaluation
---

# Experiment 006: Zero-Config Autonomous Bug Fix Engine with Self-Improving Meta-Loop
# 4. Zero-config autonomous bug fix engine with self-improving meta-loop

**Date:** 2026-03-29
**Status:** Complete
Expand Down
2 changes: 1 addition & 1 deletion 0005-agent-scoped-tools-triage/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7,7 +7,7 @@ topics:
- skills
---

# PoC FullSend — Triage Agents with scoped tools and sandbox
# 5. Agent scoped tools triage

This experiment addresses: https://github.com/fullsend-ai/fullsend/issues/101

Expand Down
4 changes: 2 additions & 2 deletions 0006-code-agent-evaluation/EXPERIMENT.md → 0006-code-agent-evaluation/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,7 +6,7 @@ topics:
- security
---

# Code Agent Evaluation Experiment
# 6. Code agent evaluation

**Date:** April 11–14, 2026
**Author:** Adam Scerra
Expand Down Expand Up @@ -853,7 +853,7 @@ V8 (hybrid — proposed in PR #189):

```
experiments/code-agent-evaluation/
├── EXPERIMENT.md # This document
├── README.md # This document
├── RECOMMENDATION.md # Action summary for PR #189
├── variants/
│ ├── VARIANTS.md # Index of all variants with git refs
Expand Down
8 changes: 4 additions & 4 deletions 0006-code-agent-evaluation/RECOMMENDATION.md
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,7 +1,7 @@
# Code Agent Recommendation — Best Configuration for PR #189

**Date:** April 14, 2026 (updated after Round 4)
**Based on:** [EXPERIMENT.md](EXPERIMENT.md) — 490+ trials across 20 scenarios, 8 variants, 4 rounds
**Based on:** [README.md](README.md) — 490+ trials across 20 scenarios, 8 variants, 4 rounds
**Target:** [PR #189: Add code agent definition and skill](https://github.com/fullsend-ai/fullsend/pull/189)
**Review feedback:** [ralphbean's preliminary review](https://github.com/fullsend-ai/fullsend/pull/189#issuecomment-1)

Expand Down Expand Up @@ -36,15 +36,15 @@ further redesign is needed.


For detailed scores by round and scenario, see
[EXPERIMENT.md sections 5–8](EXPERIMENT.md#5-round-1-results-all-variants).
[README.md sections 5–8](README.md#5-round-1-results-all-variants).

---

## Changes applied to PR #189

All priority-1 changes from the initial recommendation have been implemented
in the V8 hybrid variant, validated with 66 trials
([Round 4](EXPERIMENT.md#8-round-4-results-v8-hybrid-validation)), and
([Round 4](README.md#8-round-4-results-v8-hybrid-validation)), and
committed as [`d70dcff`](https://github.com/fullsend-ai/fullsend/commit/d70dcff).


Expand Down Expand Up @@ -110,7 +110,7 @@ structured variants)

## Cross-references

- **Full experiment:** [EXPERIMENT.md](EXPERIMENT.md)
- **Full experiment:** [README.md](README.md)
- **V8 variant (PR #189 latest):** [variants/V8-hybrid/](variants/V8-hybrid/)
- **All variants (V1–V7):** [variants/VARIANTS.md](variants/VARIANTS.md) — links to browse any variant
- **PR #189:** [Add code agent definition and skill](https://github.com/fullsend-ai/fullsend/pull/189)
Expand Down
2 changes: 1 addition & 1 deletion 0006-code-agent-evaluation/scripts/invoke-variant.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -134,7 +134,7 @@ case "${VARIANT}" in

V4)
# V4: CLAUDE.md only — infrastructure exists but V4 was excluded from
# scored results (started then stopped; see EXPERIMENT.md section 2)
# scored results (started then stopped; see README.md section 2)
VARIANT_DIR="${EXPERIMENT_ROOT}/variants/V4-claudemd-only"

# Copy CLAUDE.md to repo root
Expand Down
2 changes: 1 addition & 1 deletion 0007-github-actions-agent-runtime-mvp/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,7 +6,7 @@ topics:
- tooling
---

# Experiment: GitHub Actions Agent Runtime MVP
# 7. GitHub Actions agent runtime MVP
Comment thread
qodo-code-review[bot] marked this conversation as resolved.

## Hypothesis

Expand Down
2 changes: 1 addition & 1 deletion 0008-guardrails-eval/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,7 +6,7 @@ topics:
- evaluation
---

# Experiment: Pluggable Guardrails Evaluation
# 8. Guardrails evaluation

Evaluates prompt injection scanning tools as pre-LLM guardrails for fullsend's autonomous SDLC pipeline. Extends the [Model Armor experiment](../0012-model-armor-vs-agent-triage/) (PR #117) with additional attack payloads and alternative scanning backends.

Expand Down
2 changes: 1 addition & 1 deletion 0009-hermes-security-patterns/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,7 +6,7 @@ topics:
- tooling
---

# Experiment: Hermes-Inspired Security Patterns for Fullsend
# 9. Hermes-inspired security patterns

Evaluates security patterns from [Hermes Agent](https://github.com/NousResearch/hermes-agent) for integration into fullsend's autonomous SDLC pipeline. Tests two integration strategies: static file scanning via [Tirith](https://github.com/sheeki03/tirith) CLI and runtime SSRF protection via a Claude Code PreToolUse hook.

Expand Down
2 changes: 1 addition & 1 deletion 0010-host-side-api-server/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,7 +6,7 @@ topics:
- tooling
---

# Experiment: Host-Side API Server for Sandboxed Agents
# 10. Host-side API server for sandboxed agents

Tracking issue: [fullsend-ai/experiments#25](https://github.com/fullsend-ai/experiments/issues/25)

Expand Down
2 changes: 1 addition & 1 deletion 0011-integration-service-design-drift/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,7 +6,7 @@ topics:
- tooling
---

# Experiment: Integration Service Design Doc Drift Analysis
# 11. Integration Service design doc drift analysis

**Date:** 2026-03-20

Expand Down
2 changes: 1 addition & 1 deletion 0012-model-armor-vs-agent-triage/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,7 +6,7 @@ topics:
- evaluation
---

# Experiment: Model Armor vs AI Agent Triage — Prompt Injection
# 12. Model Armor vs AI agent triage

## Hypothesis

Expand Down
2 changes: 1 addition & 1 deletion 0013-openshell-policy-bypass/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,7 +6,7 @@ topics:
- sandbox
---

# OpenShell Binary Policy Bypass Experiment
# 13. OpenShell policy bypass

This experiment tests whether OpenShell's binary-level network policy
restrictions can be bypassed by an agent. It answers two questions:
Expand Down
2 changes: 1 addition & 1 deletion 0014-openshell-sandbox-evaluation.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,7 +6,7 @@ topics:
- evaluation
---

# Experiment: OpenShell Sandbox Evaluation
# 14. OpenShell sandbox evaluation

**Date:** 2026-03-27
**Status:** Complete
Expand Down
2 changes: 1 addition & 1 deletion 0015-prompt-injection-defense/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,7 +6,7 @@ topics:
- evaluation
---

# Experiment 002: Prompt Injection Defense-in-Depth
# 15. Prompt injection defense-in-depth

## Hypothesis

Expand Down
2 changes: 1 addition & 1 deletion 0016-promptfoo-eval/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,7 +6,7 @@ topics:
- tooling
---

# Experiment 004: Promptfoo for Agent Evaluation in CI
# 16. Promptfoo for agent evaluation in CI

**Date:** 2026-03-20
**Status:** Complete
Expand Down
2 changes: 1 addition & 1 deletion 0017-reasoning-monitor/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,7 +6,7 @@ topics:
- evaluation
---

# Experiment 005: Reasoning Monitor Agent for Prompt Injection Detection
# 17. Reasoning monitor agent for prompt injection detection

**Status:** Implementation complete, results pending
**Issue:** [fullsend #174](https://github.com/konflux-ci/fullsend/issues/174)
Expand Down
2 changes: 1 addition & 1 deletion 0018-runner-hello-world/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,7 +6,7 @@ topics:
- sandbox
---

# Runner Hello World Experiment
# 18. Runner hello world

A minimal end-to-end test of the `fullsend run` CLI. It provisions an OpenShell sandbox from a pre-built container image, runs a Claude Code agent inside it, extracts output, and validates the result with a retry loop.

Expand Down
2 changes: 1 addition & 1 deletion 0019-skills/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5,6 +5,6 @@ topics:
- skills
---

# Experiment: Skills
# 19. Skills

Experiments with skill authoring and consumption patterns.
2 changes: 1 addition & 1 deletion 0020-target-repo-skills/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,7 +6,7 @@ topics:
- tooling
---

# Experiment: Target Repository Skills in Triage
# 20. Target repository skills in triage

## Hypothesis

Expand Down
2 changes: 1 addition & 1 deletion 0021-tool-scoping/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,7 +6,7 @@ topics:
- tooling
---

# ADR 0022 Tool Scoping Experiment
# 21. Tool scoping

Evidence for
[ADR 0022](https://github.com/fullsend-ai/fullsend-adr-tools/blob/main/docs/ADRs/0022-allowed-and-disallowed-tools-for-agents.md).
Expand Down
2 changes: 1 addition & 1 deletion 0022-claude-github-app-auth/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,7 +6,7 @@ topics:
- security
---

# PoC FullSend #67
# 22. Claude GitHub App auth

This experiment folder addresses: https://github.com/konflux-ci/fullsend/issues/67

Expand Down
14 changes: 11 additions & 3 deletions review-cache-experiment/README.md → 0023-review-cache/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,4 +1,12 @@
# Review Cache - Publication Policy Architecture
---
title: "23. Review cache publication policy"
status: Concluded
topics:
- review
- deduplication
---

# 23. Review cache publication policy

## Problem

Expand Down Expand Up @@ -204,7 +212,7 @@ Total integration: ~40 lines of bash
## File Structure

```
review-cache-experiment/
0023-review-cache/
├── scripts/ # Production code (post-scripts only)
│ ├── store.py # Layer 1: SQLite storage
│ ├── save.py # Layer 2: Deduplication
Expand Down Expand Up @@ -356,4 +364,4 @@ To validate readiness for production integration, the experiment needs real data

**Deduplication accuracy:** Needs real-world validation before production use. The 100% success rate on synthetic data validates the mechanism works, but doesn't yet prove it handles the messy reality of production code reviews.

**Recommendation:** Run real data validation to measure actual dedup rates and identify edge cases before committing to SQLite cache as the production solution.
**Recommendation:** Run real data validation to measure actual dedup rates and identify edge cases before committing to SQLite cache as the production solution.
0 ...e-experiment/scripts/extract_functions.sh → ...review-cache/scripts/extract_functions.sh
View file
Open in desktop
File renamed without changes.
Loading
Loading