Detect Zero Iteration Failures
HKUDS/OpenSpace
Identify and handle pre-execution agent failures occurring before any iterations or tool usage
Detecting whether agent iterations are converging toward a stable solution or hitting a ceiling.
$ npx skills add hashgraph-online/awesome-codex-plugins --skill convergence-monitoring -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install hashgraph-online/awesome-codex-plugins convergence-monitoring --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
$ git clone --depth 1 https://github.com/hashgraph-online/awesome-codex-plugins.git skills-src && mkdir -p .claude/skills && cp -r skills-src/plugins/JuliusBrussee/blueprint/skills/convergence-monitoring .claude/skills/convergence-monitoring && rm -rf skills-srcUse ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.
Claude Code skills documentation · loads skills from .claude/skills/
Install the "convergence-monitoring" agent skill from https://github.com/hashgraph-online/awesome-codex-plugins/tree/main/plugins/JuliusBrussee/blueprint/skills/convergence-monitoring into .claude/skills/convergence-monitoring/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "convergence-monitoring", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$skill-installer install https://github.com/hashgraph-online/awesome-codex-plugins/tree/main/plugins/JuliusBrussee/blueprint/skills/convergence-monitoringType this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.
$ npx skills add hashgraph-online/awesome-codex-plugins --skill convergence-monitoring -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install hashgraph-online/awesome-codex-plugins convergence-monitoring --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/hashgraph-online/awesome-codex-plugins.git skills-src && mkdir -p .agents/skills && cp -r skills-src/plugins/JuliusBrussee/blueprint/skills/convergence-monitoring .agents/skills/convergence-monitoring && rm -rf skills-srcUse ~/.agents/skills/ instead of .agents/skills for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "convergence-monitoring" agent skill from https://github.com/hashgraph-online/awesome-codex-plugins/tree/main/plugins/JuliusBrussee/blueprint/skills/convergence-monitoring into .agents/skills/convergence-monitoring/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "convergence-monitoring", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add hashgraph-online/awesome-codex-plugins --skill convergence-monitoring -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install hashgraph-online/awesome-codex-plugins convergence-monitoring --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/hashgraph-online/awesome-codex-plugins.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/plugins/JuliusBrussee/blueprint/skills/convergence-monitoring .cursor/skills/convergence-monitoring && rm -rf skills-srcUse ~/.cursor/skills/ instead of .cursor/skills for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "convergence-monitoring" agent skill from https://github.com/hashgraph-online/awesome-codex-plugins/tree/main/plugins/JuliusBrussee/blueprint/skills/convergence-monitoring into .cursor/skills/convergence-monitoring/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "convergence-monitoring", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gemini skills install https://github.com/hashgraph-online/awesome-codex-plugins.git --path plugins/JuliusBrussee/blueprint/skills/convergence-monitoring--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.
$ npx skills add hashgraph-online/awesome-codex-plugins --skill convergence-monitoring -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install hashgraph-online/awesome-codex-plugins convergence-monitoring --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/hashgraph-online/awesome-codex-plugins.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/plugins/JuliusBrussee/blueprint/skills/convergence-monitoring .gemini/skills/convergence-monitoring && rm -rf skills-srcUse ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "convergence-monitoring" agent skill from https://github.com/hashgraph-online/awesome-codex-plugins/tree/main/plugins/JuliusBrussee/blueprint/skills/convergence-monitoring into .gemini/skills/convergence-monitoring/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "convergence-monitoring", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install hashgraph-online/awesome-codex-plugins convergence-monitoringInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add hashgraph-online/awesome-codex-plugins --skill convergence-monitoring -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
$ git clone --depth 1 https://github.com/hashgraph-online/awesome-codex-plugins.git skills-src && mkdir -p .github/skills && cp -r skills-src/plugins/JuliusBrussee/blueprint/skills/convergence-monitoring .github/skills/convergence-monitoring && rm -rf skills-srcUse ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "convergence-monitoring" agent skill from https://github.com/hashgraph-online/awesome-codex-plugins/tree/main/plugins/JuliusBrussee/blueprint/skills/convergence-monitoring into .github/skills/convergence-monitoring/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "convergence-monitoring", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add hashgraph-online/awesome-codex-plugins --skill convergence-monitoring -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install hashgraph-online/awesome-codex-plugins convergence-monitoring --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/hashgraph-online/awesome-codex-plugins.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/plugins/JuliusBrussee/blueprint/skills/convergence-monitoring .opencode/skills/convergence-monitoring && rm -rf skills-srcUse ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "convergence-monitoring" agent skill from https://github.com/hashgraph-online/awesome-codex-plugins/tree/main/plugins/JuliusBrussee/blueprint/skills/convergence-monitoring into .opencode/skills/convergence-monitoring/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "convergence-monitoring", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
convergence-monitoringDetecting whether agent iterations are converging toward a stable solution or hitting a ceiling.
Convergence Monitoring is an agent skill from hashgraph-online/awesome-codex-plugins. Detecting whether agent iterations are converging toward a stable solution or hitting a ceiling. Covers convergence signals, ceiling detection, non-convergence diagnosis, test pass rate as a convergence metric, and forward progress tracking for large projects. Trigger phrases: "convergence", "is the agent converging", "ceiling detection", "when to stop iterating", "diminishing returns"
Its SKILL.md is about 4.2k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.
The repository describes itself as: A curated list of awesome OpenAI Codex / ChatGPT plugins, skills, and resources. The 1 Codex Marketplace. See live plugins at: https://hol.org/plugins/best-codex-plugins. The licence is Apache-2.0.
9 steps, taken from the step headings in SKILL.md.
Read from SKILL.md and the folder at commit 3e1456a. It shows what the files ask for, not the result of running them.
Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.
From allowed-tools in the SKILL.md frontmatter.
Shell commands in SKILL.md call:
gitFrom the folder's file list and the shell code blocks in SKILL.md.
No URLs in SKILL.md. Its commands use git, which can reach the network depending on how they are called.
From URLs in SKILL.md, links to its own repository left out.
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
Convergence Monitoring loads about 4.2k tokens when it runs. Until then it costs about 103 tokens; SKILL.md has 1,281 words of instructions outside code blocks.
Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.
The automated check found no risky patterns in SKILL.md.
Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.
The full file from hashgraph-online/awesome-codex-plugins at commit 3e1456a, republished under its Apache-2.0 licence (© hashgraph-online). 1,281 words, ~4,209 tokens.
.claude/skills/convergence-monitoring/SKILL.md (or your agent's skills folder).Convergence monitoring answers the most important question in iterative AI development: when should you stop iterating? The answer is not a fixed number of iterations or a time limit -- it is convergence. Convergence means the agent's output is stabilizing; each iteration produces fewer and smaller changes than the last.
Core insight: You don't need a zero-diff -- you need the remaining modifications to be inconsequential.
Convergence appears as a rapid, consistent decline in the volume of changes from one iteration to the next:
Iteration 1: ████████████████████████████████████████ 300 lines changed
Iteration 2: ████████████████ 120 lines changed
Iteration 3: ██████ 40 lines changed
Iteration 4: ██ 10 lines changed (cosmetic only)
^--- Convergence reached: the diff shrinks each pass until only cosmetic changes remain| Signal | What It Means |
|---|---|
| Lines changed decreasing exponentially | Each iteration makes roughly half the changes of the previous one |
| Changes become trivial | Remaining changes are formatting, comments, imports -- not behavior |
| Tests stabilize | Test count stops increasing; pass rate approaches 100% |
| No new files created | The architecture has settled; only existing files are modified |
| Impl tracking updates shrink | Implementation tracking changes are status updates, not new findings |
| Completion signal emitted | Agent determines all exit criteria are met |
# Check lines changed per iteration
git log --oneline --stat
# Iteration 5: trivial changes
abc1234 Iteration 5: formatting and comment fixes
3 files changed, 8 insertions(+), 6 deletions(-)
# Iteration 4: minor adjustments
def5678 Iteration 4: edge case handling
5 files changed, 22 insertions(+), 8 deletions(-)
# Iteration 3: moderate changes
ghi9012 Iteration 3: complete API integration
12 files changed, 85 insertions(+), 31 deletions(-)
# Iteration 2: significant changes
jkl3456 Iteration 2: implement core features
18 files changed, 156 insertions(+), 42 deletions(-)
# Iteration 1: major initial work
mno7890 Iteration 1: initial implementation
25 files changed, 312 insertions(+), 15 deletions(-)A ceiling is when the agent cannot make further progress due to external constraints. Like convergence, it produces small diffs -- but for fundamentally different reasons.
Convergence: Agent is DONE -> small diffs because work is complete
Ceiling: Agent is STUCK -> small diffs because agent cannot proceed| Cause | Example | How to Detect |
|---|---|---|
| Missing dependency | API not available, library not installed | Agent logs errors about unavailable resources |
| Ambiguous spec | Requirement can be interpreted multiple ways | Agent oscillates between implementations |
| Tooling limitation | Build tool does not support needed feature | Agent tries workarounds that do not converge |
| External service | Test requires network access, external API | Tests fail with connection/timeout errors |
| Context window exhaustion | Codebase too large for one session | Agent loses track of earlier work |
| Permission boundary | Agent cannot access needed files or systems | Repeated permission errors in logs |
| Dimension | Convergence (work is finishing) | Ceiling (work is stuck) |
|---|---|---|
| Size of diffs | Shrinking steadily toward zero | Staying small but not trending down |
| Nature of changes | Cosmetic -- whitespace, comments, naming | Functional but going in circles |
| Test results | Pass rate climbing toward full coverage | Pass rate plateaued below target |
| Agent stance | Wrapping up, marking exit criteria done | Retrying the same strategies repeatedly |
| Tracking status | Tasks moving to DONE | BLOCKED items piling up |
| Recommended action | Declare done, move to next phase | Diagnose the obstacle, resolve it, then continue |
Check 1: Are tests passing?
YES, and improving -> Convergence
NO, stuck at same failures -> Ceiling
Check 2: Is the agent trying new approaches?
NO, just polishing -> Convergence
YES, but they all fail similarly -> Ceiling
Check 3: Are there BLOCKED tasks in impl tracking?
NO -> Convergence
YES -> Ceiling (read the blockers)
Check 4: Is the agent producing meaningful error messages?
NO, just minor changes -> Convergence
YES, about dependencies/tools/access -> CeilingNon-convergence means the agent is making changes, but they are NOT decreasing. The system is not stabilizing.
Non-convergence:
Iteration 1: ████████████████████████████████████████ 250 lines changed
Iteration 2: ██████████████████████████████████████ 230 lines changed
Iteration 3: ████████████████████████████████████████ 260 lines changed
Iteration 4: ██████████████████████████████████ 220 lines changed
^--- NOT converging: changes are flat/oscillating| Root Cause | Symptom | Fix |
|---|---|---|
| Fuzzy specs | Agent interprets requirements differently each iteration | Make specs more precise; add concrete acceptance criteria |
| Weak validation | Agent cannot verify correctness, so it keeps changing things | Add build/test/lint gates; strengthen acceptance criteria |
| Fighting sub-agents | Multiple agents change the same code in conflicting ways | Add file ownership tables; dispatch subagents via the Agent tool |
| Contradictory requirements | Spec A says X, spec B says not-X | Resolve contradictions in specs; add explicit priority/precedence |
| Missing exit criteria | Agent does not know when it is done | Add explicit exit criteria checklists and completion signals |
| Over-broad scope | Too much work for one prompt/iteration | Split into smaller, focused prompts with clear boundaries |
| Unstable dependencies | External library or API keeps changing | Pin dependencies; mock external services in tests |
When the loop isn't stabilizing, the problem is upstream -- fix the specifications, validation, or coordination rather than adding more passes.
Running more iterations when the system is not converging wastes time and compute. Instead:
Test pass rate is the most reliable quantitative convergence signal. Track these metrics:
| Iteration | Tests | Pass | Fail | Skip | Pass Rate | Delta |
|-----------|-------|------|------|------|-----------|-------|
| 1 | 45 | 30 | 15 | 0 | 66.7% | -- |
| 2 | 62 | 50 | 12 | 0 | 80.6% | +13.9 |
| 3 | 78 | 70 | 8 | 0 | 89.7% | +9.1 |
| 4 | 85 | 82 | 3 | 0 | 96.5% | +6.8 |
| 5 | 88 | 87 | 1 | 0 | 98.9% | +2.4 || Pattern | Meaning | Action |
|---|---|---|
| Test count increasing | Agent is adding coverage | Good -- system is maturing |
| Pass rate approaching 100% | Implementation matches specs | Good -- approaching convergence |
| Fewer failures per iteration | Each pass fixes more than it breaks | Good -- healthy convergence |
| Pass rate plateaus < 100% | Some tests consistently fail | Ceiling -- investigate failing tests |
| Test count decreasing | Agent is deleting tests | Bad -- investigate why; may be deleting inconvenient tests |
| Pass rate oscillating | Fixes in one area break another | Non-convergence -- check for conflicting specs |
# After each iteration, check convergence signals
echo "=== Convergence Check ==="
# 1. Lines changed (should be decreasing)
git diff --stat HEAD~1
# 2. Test results (should be improving)
{TEST_COMMAND} 2>&1 | tail -5
# 3. Build health (should always pass)
{BUILD_COMMAND} 2>&1 | tail -3
# 4. Files changed (should be decreasing)
git diff --name-only HEAD~1 | wc -lFor large projects where full convergence takes many iterations, track forward progress toward eventual convergence.
The percentage of spec requirements with passing tests:
Spec Requirements Coverage:
spec-auth.md: ██████████████████████████████████████ 95% (19/20 requirements)
spec-data.md: ████████████████████████████████ 80% (16/20 requirements)
spec-ui.md: ██████████████████████ 55% (11/20 requirements)
spec-api.md: ████████████████████████████ 70% (14/20 requirements)
─────────────────────────────────────────────────────
Overall: ████████████████████████████ 75% (60/80 requirements)| Metric | Healthy Trend | Unhealthy Trend |
|---|---|---|
| Requirements with passing tests | Increasing each iteration | Flat or decreasing |
| Total test count | Increasing | Flat or decreasing |
| DONE tasks in impl tracking | Increasing | Flat with BLOCKED tasks growing |
| Open issues | Decreasing | Increasing or flat |
| Dead ends documented | Increasing slightly (learning) | Exploding (thrashing) |
Track how much progress each iteration makes:
| Iteration | Requirements Met | New This Iteration | Velocity |
|-----------|-----------------|-------------------|----------|
| 1 | 15/80 | 15 | 15 |
| 2 | 30/80 | 15 | 15 |
| 3 | 48/80 | 18 | 18 |
| 4 | 60/80 | 12 | 12 |
| 5 | 68/80 | 8 | 8 |
| 6 | 73/80 | 5 | 5 |
| 7 | 76/80 | 3 | 3 |Velocity should decrease over time (easy requirements first, hard ones last), but should never hit zero. Zero velocity = ceiling.
Stop the iteration loop when ANY of these are true:
<all-tasks-complete>[ ] in the exit criteria checklist is [x]Continue iterating when ALL of these are true:
Pause and investigate when ANY of these are true:
+------------------------------------------------------+
| Convergence Dashboard |
+------------------------------------------------------+
| Iteration: 4/10 |
| Lines changed: 45 (prev: 112, trend: decreasing) |
| Files changed: 3 (prev: 8, trend: decreasing) |
| Test pass rate: 94.2% (prev: 87.1%, trend: up) |
| Tests: 82 total (prev: 75, trend: up) |
| BLOCKED tasks: 0 (prev: 1, trend: down) |
| Status: CONVERGING |
+------------------------------------------------------+# Quick convergence check after each iteration
echo "--- Lines changed ---"
git diff --stat HEAD~1 | tail -1
echo "--- Files changed ---"
git diff --name-only HEAD~1 | wc -l
echo "--- Test results ---"
{TEST_COMMAND} --summary 2>&1 | tail -3
echo "--- Impl tracking status ---"
grep -c "BLOCKED\|IN_PROGRESS\|TODO\|DONE" context/impl/impl-*.mdSet up alerts for non-convergence signals:
| Alert | Trigger | Action |
|---|---|---|
| Oscillation | Lines changed increased vs previous iteration | Pause; check for conflicting changes |
| Stall | Lines changed < 5 but tests still failing | Pause; likely a ceiling |
| Regression | Test pass rate decreased | Pause; investigate what broke |
| Runaway | Lines changed > 500 for 3+ iterations | Pause; scope may be too broad |
When you detect non-convergence, follow this recovery process:
Do not keep running. More iterations will not help.
## Non-Convergence Diagnosis
### Symptoms
- [ ] Changes are flat (not decreasing)
- [ ] Changes are oscillating (up-down-up-down)
- [ ] Agent is retrying failed approaches
- [ ] Tests are oscillating (passing then failing)
- [ ] Multiple agents changing the same files
### Root Cause Analysis
1. Check specs: Are requirements clear and unambiguous?
2. Check validation: Can the agent verify correctness?
3. Check file ownership: Are agents conflicting?
4. Check scope: Is the prompt trying to do too much?
5. Check dependencies: Are external resources available?| Root Cause | Fix |
|---|---|
| Fuzzy specs | Rewrite ambiguous requirements with concrete acceptance criteria |
| Weak validation | Add build/test/lint gates to the prompt |
| File conflicts | Add file ownership tables; dispatch subagents via the Agent tool |
| Over-broad scope | Split into smaller prompts; reduce concurrent agents |
| External dependency | Mock the dependency; or resolve it before resuming |
After fixing the root cause, resume from where you stopped. Do NOT restart from scratch -- git history preserves all progress.
# Resume with the same prompt, possibly fewer remaining iterations
iteration-loop context/prompts/003-generate-impl-from-plans.md -n 5 -t 1hRevision directly improves convergence by making specs more complete:
Without revision:
Iteration 1: 200 lines, 5 manual fixes -> specs unchanged
Iteration 2: 180 lines, 4 manual fixes -> specs unchanged
Iteration 3: 170 lines, 4 manual fixes -> NOT converging
With revision:
Iteration 1: 200 lines, 5 manual fixes -> specs updated with 5 new requirements
Iteration 2: 100 lines, 2 manual fixes -> specs updated with 2 new requirements
Iteration 3: 50 lines, 0 manual fixes -> CONVERGINGFrequent manual fixes without revision = non-convergence. The iteration loop keeps producing the same bugs because nothing in the specs prevents them.
references/convergence-patterns.md for the complete convergence pattern catalog with examples.ck:revision skill for how tracing bugs to specs improves convergence.ck:prompt-pipeline skill for designing prompts with proper exit criteria and completion signals.ck:validation-first skill for building validation gates that provide convergence signals.ck:impl-tracking skill for tracking progress and detecting ceiling conditions.© hashgraph-online, Apache-2.0. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file
Just SKILL.md in plugins/JuliusBrussee/blueprint/skills/convergence-monitoring of hashgraph-online/awesome-codex-plugins.
Open the folder on GitHubat commit 3e1456a
Convergence Monitoring next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| Convergence Monitoring this skillhashgraph-online/awesome-codex-plugins | 1.3k | — | ~4.2k | Automated safety check: Pass | Apache-2.0 | |
| Detect Zero Iteration FailuresHKUDS/OpenSpace | 7.8k | — | ~1.3k | Automated safety check: Pass | MIT | |
| Detect Zero Iteration FailureHKUDS/OpenSpace | 7.8k | — | ~477 | Automated safety check: Pass | MIT | |
| Threat Detectionalirezarezvani/claude-skills | 28k | — | ~3.5k | Automated safety check: Pass | MIT | |
| Resemble Detectgithub/awesome-copilot | 40k | 3 repos | ~4.1k | Automated safety check: Pass | Apache-2.0 | |
| Pii Detectruvnet/ruflo | 74k | — | ~350 | Automated safety check: Notes | MIT |
HKUDS/OpenSpace
Identify and handle pre-execution agent failures occurring before any iterations or tool usage
HKUDS/OpenSpace
Identify and classify agent failures occurring before execution starts
alirezarezvani/claude-skills
A skill your agent uses when hunting for threats in an environment, analyzing IOCs, or detecting behavioral anomalies in telemetry.
github/awesome-copilot
Deepfake detection and media safety — detect AI-generated audio, images, video, and text, trace synthesis sources, apply watermarks, verify speaker identity, and analyze media intelligence using…
ruvnet/ruflo
Detect and flag personally identifiable information (PII) in text, code, and configurations.
mukul975/Anthropic-Cybersecurity-Skills
Detect API enumeration attacks (BOLA/IDOR, OWASP API1:2023) by writing SIEM detection rules that flag sequential or UUID identifier iteration, parameter tampering, and mixed 200/401/403 response…
hashgraph-online/awesome-codex-plugins
Create original anime-style reaction stickers as looping GIFs and MP4 previews, using generated character pose sheets and timed key poses.
hashgraph-online/awesome-codex-plugins
Manage and query Calibre libraries with the calibredb CLI (local paths or Calibre Content server URLs).
hashgraph-online/awesome-codex-plugins
A skill your agent uses when adding, changing, testing, or debugging Rust HTTP APIs and services, especially when Codex needs black-box integration tests, random-port app startup, real database test…
hashgraph-online/awesome-codex-plugins
Make a studio's game look like something at build time — a cover from a real frame of the game (free), painted covers, backdrops, textures and character plates from image models through the…
hashgraph-online/awesome-codex-plugins
Use CALL-E from Codex through the calle CLI. An agent skill from hashgraph-online/awesome-codex-plugins.
hashgraph-online/awesome-codex-plugins
Balance game difficulty, resources, rewards, probability, progression, economies, and dominant strategies.
Detecting whether agent iterations are converging toward a stable solution or hitting a ceiling. Convergence Monitoring is an agent skill from hashgraph-online/awesome-codex-plugins. Detecting whether agent iterations are converging toward a stable solution or hitting a ceiling.
Convergence Monitoring fits situations like: phrases: convergence; is the agent converging; ceiling detection; to stop iterating.
Run `npx skills add hashgraph-online/awesome-codex-plugins --skill convergence-monitoring -a claude-code`. Or copy the skill folder (plugins/JuliusBrussee/blueprint/skills/convergence-monitoring in hashgraph-online/awesome-codex-plugins) into .claude/skills/convergence-monitoring in your project. Claude Code loads it when a task matches its description.
Run `npx skills add hashgraph-online/awesome-codex-plugins --skill convergence-monitoring -a codex`. Or copy the skill folder (plugins/JuliusBrussee/blueprint/skills/convergence-monitoring in hashgraph-online/awesome-codex-plugins) into .agents/skills/convergence-monitoring in your project. Codex loads it when a task matches its description.
Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add hashgraph-online/awesome-codex-plugins --skill convergence-monitoring -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/convergence-monitoring, .gemini/skills/convergence-monitoring, .github/skills/convergence-monitoring and .opencode/skills/convergence-monitoring in your project.
Going by SKILL.md and its folder, Convergence Monitoring needs the command-line tools its instructions call (git).
SKILL.md contains no URLs. Its commands use git, which can reach the network depending on how they are called. This is read from the text; nothing was executed.
Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.
Convergence Monitoring is published under the Apache-2.0 licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.
About 4.2k tokens (SKILL.md is roughly 17k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.
Skills that share tags, products or a category with Convergence Monitoring: Detect Zero Iteration Failures (HKUDS/OpenSpace, 7.8k stars), Detect Zero Iteration Failure (HKUDS/OpenSpace, 7.8k stars), Threat Detection (alirezarezvani/claude-skills, 28k stars) and Resemble Detect (github/awesome-copilot, 40k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
hashgraph-online (a GitHub organization) maintains it in hashgraph-online/awesome-codex-plugins, which has 1,267 GitHub stars. The repository holds 716 skills in this directory. The repository was last updated on October 10, 2026.
Source: hashgraph-online/awesome-codex-plugins on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.